File:February 2023 Wikimedia Enterprise API community conversation meeting.webm

0:04
hello good evening or good day my name is Liam Wyatt known as user Wittylama
0:10
this is the February 2023
0:15
office hours community call for the Wikimedia Enterprise project this is in
0:21
the immediate week of the publication of
0:29
the inaugural financial report of this project and also a product update
0:36
for some future software changes that are being introduced and this is the call to for anyone in the Wikipedia
0:43
Community who has any questions and wishes to ask them live the video is
0:48
being recorded and will be uploaded to Wikipedia comments as a result this session follows in the scope of the uh
0:56
saga space policy of the Wikimedia Foundation anyone can obviously have their camera
1:02
off or on or write comments in the chat as they like I will put a link in the chat and also
1:09
in the Wikipedia Commons file the link to the financial report
1:16
and I will try and keep notes of time code notes for when each question is listed so it's easy for people to come
1:22
back afterwards uh introducing the people from the Wikimedia Enterprise team in the room
1:28
perhaps if you could just uh introduce yourselves briefly those of you who who wish to
1:34
study with Lane sure and hi folks I am Lane Becker I am
1:39
the head of the Wikimedia Enterprise project at the foundation
1:45
um my primary focus is working with many of the folks here on the operational aspects of the business that includes
1:51
sort of figuring fitting it into the sort of the workings of the larger Foundation including financially so
1:57
quite involved alongside Liam with the creation of the financial reporting
2:03
uh uh and then just generally any aspect of business operations or sales
2:10
I'll pass it to Amy hi everybody
2:16
um I'm Amy Muller and I am the manager of customer success and support for Wikimedia Enterprise and that is exactly
2:24
what it sounds like another cutter
2:30
yes hi everybody I'm really relatively new Joiner in the team I'm a software
2:36
engineering the Greek Community Enterprise team and that's it for me Chuck Maybe
2:44
hello uh my name is Chuck and I am product marketing primarily and also
2:49
helping and around product itself and thank you
2:58
hello everyone my name is and I'm uh not operations manager on the
3:06
Enterprise team can I should also point out we have Dennis Bartel here uh who's not directly
3:15
with the Enterprise team but is a Wikipedia Foundation team member responsible for
3:21
foreign yeah I'm working with the movement strategy and governance team and I'm
3:27
here to support and assist in cases yeah and when it comes around with the German
3:32
community maybe to help and clarify um wordings and similar
3:37
that's what I pay for so we do actually in this team uh present here uh speak obviously English
3:45
French German Italian Portuguese possibly other language Dutch
3:52
um so if anyone has questions in those languages they can ask them natively if
3:58
they would like that all being said I passed the floor to uh anyone who has a burning question
4:06
of already in their mind from the financial report
4:11
otherwise I can describe the financial report verbally for those who have not read it yet or
4:17
have not seen it yet but I want to stop talking myself if people who'd come to
4:22
this meeting with a idea of a question in mind already
4:28
foreign
4:38
in that case I will describe what we have done here in this
4:44
report and hopefully that will generate some concerns or questions we do have
4:50
some questions pre-listed from the um mailing list and from the talk page so I
4:58
want to reiterate those uh later but I want to make sure people who are physically here in this call get
5:05
the first opportunity so please raise your hand if you have a question whenever you feel like it otherwise the
5:11
purpose of the financial report was that the Wikipedia Enterprise project is operated by a completely
5:19
owned commercial subsidiary organization of the Burkina Foundation a non-profit
5:25
organization both in America and as a commercial activity
5:32
uh it is has its own Financial requirements its own Financial purpose
5:39
to make a profit to give money back to the Wikimedia movement but
5:45
because that is entirely owned by the Wikimedia Foundation a non-profit organization there is no standard
5:53
required form for that financial report there is for the Wikipedia foundation
5:59
for non-profit organizations that is called the form 990 form 990 which is
6:05
published by the Wikipedia Foundation every year but this thing the committee Enterprise does not have an equivalent
6:11
so we wanted to make sure we had something published that was clear and
6:17
just for this project not hidden or unclear in the larger
6:24
Wikipedia Foundation report for the needs of the tax department in
6:31
America the IRS and for other legal and financial accountability requirements
6:36
all of this information is in the form 990 but that is not an easy thing to
6:42
read especially if you are not an American accountant uh it doesn't does not actually help to explain anything so
6:49
this report is written as a narrative it describes rather than just lists numbers
6:55
in a and clear structure the key messages
7:01
really are that we are approximately covering costs uh some months were
7:09
slightly above the cost of that month some months the revenue was slightly below the cost of that month but really
7:15
within a quite a narrow percentage range so we are at the moment
7:21
equal for Revenue per month and expense per month um this we hope will change in the
7:30
positive next year but we'll uh we can't promise that but that is the expectation
7:38
the uh taxation is a common uh question or
7:44
common request the short answer for the quick for the issue of Taxation is we did not pay any
7:50
tax because we were not yet profitable over the course of the entire year there is a much longer answer for that which
7:56
covers state and federal uh different uh rules like that but the the simple
8:03
answer is you don't pay tax if you're not yet profitable and we are not yet profitable because we have only been
8:09
existing for a year as a commercial activity the
8:15
organization itself the LLC limited liability company is entirely owned by the Wikimedia
8:23
foundation and all of us myself Lane the others in this team here are employees of public media Foundation it exists uh
8:31
entirely to be the legal thing that signs contracts with
8:37
commercial organizations and owns the risk of the legal risk of that promise
8:43
this ensures that the Wikimedia Foundation itself cannot be sued and
8:49
lose a billion dollars if we do something wrong that it's illegal protection uh it is not
8:57
it makes zero difference from a financial perspective and it makes zero difference from a reporting
9:05
transparency perspective which state that of America that LLC is registered
9:12
in this is a common question it is registered in the state of
9:18
Delaware because that is the state where most of corporate law in America exists
9:25
and therefore that is where contracts are most well understood to their
9:30
meaning and their implications if you need to go to court
9:35
it's not a lawyers like to know in advance what words mean
9:42
that is a summary of the of the financial section I have not touched the product section we can come back to that
9:48
there was a question from the mailing list but I do see a question in the chat
9:54
about profitability uh yeah okay you could you would feel
9:59
free to answer yourself directly okay good so yeah I have a question
10:06
about the prophet of her base
10:11
profitability so now it's all right so yeah because I think if you have none if
10:20
you offer your service for free for more than a trial use or some small
10:27
use I think then it is necessary that that
10:33
you have a kind of compensation for that or from my point of view it would be
10:38
great if then the Wikipedia Foundation but pay for these use cases for example for
10:46
the the things or services that are offered to the internet archive amount
10:52
to the Wikimedia LLC to make it transparent because I have looked at the form 990 and other then there is a
11:01
section about contributions to organizations and there is a long list
11:06
of different organizations and what the amount they receive from the
11:12
Wikimedia Foundation to support the work and I have seen that there's also
11:17
possible to to use different methods to evaluate the amount and so
11:26
from my point of view at least I think it would be possible also to evaluate the
11:32
contribution given to the internet archive for using the Wikimedia
11:39
Enterprise API yeah and I wish that this is published because I think then the
11:47
Wikimedia lse would be profitable because why at least from my it's a
11:53
speculative thing I can't guarantee that but I think they use it enough as that it
12:01
would be then profitable
12:08
I can I think either lane or Amy as the
12:13
customer service support manager uh might have different angles on this answer themselves I see ladies thinking
12:22
thinking hard there uh my immediate response to this is it I I think it
12:30
would be technically possible to account in a financial sense for the value of
12:35
the service provided to hex external XYZ external organization
12:43
uh but and that might fall under an accounting
12:49
law or an accounting rule from the finance department the fact that they
12:55
have not counted for this currently means that is
13:01
not a requirement otherwise they would do it there are various weird accounting
13:08
laws I've never heard of that because I'm not an accountant that we have to follow
13:13
um so the fact that three contracts are not formally counted as
13:21
a cost means that it is not an accounting standard to do it that way it might be
13:28
possible to describe uh in our um
13:34
report the the the kinds of usages but it is
13:41
important to also note that we cannot promise or just announce on
13:48
behalf of anyone else what they are doing with Wikimedia content without
13:54
their permission uh in exactly the same way as it is the
14:00
privacy of a reader that you can read whatever Wikipedia
14:05
article you want and the Wikimedia Foundation is not going to publish your
14:11
data without your commission you can write a blog post if you want we will
14:18
write and this is uh Chuck his uh work right uh blog posts and stories case
14:25
studies about how different customers free or or paid are using their content
14:31
but it's not in the context of a financial accounting regulation but a
14:37
use case uh from the lane can you elaborate yeah
14:44
well I was just going to say I actually don't know I just want to be upfront I
14:50
don't know how our finances will appear like how our part of the look we were
14:57
working on our own Financial transparency report to be specific about our finances but I do not yet know how
15:05
they are going to be accounted for on the form 990 that the foundation puts
15:10
out so it's unclear to me the degree to which it will be sort of separate or it will be Blended in it's certainly
15:17
something that we can try and understand better I think Liam in advance about actually happening but that's just as
15:24
one of the things that your question made me realize is I'm actually on because we tend not to
15:30
we have tended to focus mostly on just sort of making sure that we are clear about where the money is coming from and
15:36
how it is coming in to us we have been because it is out of our purview to say how the money gets spent that's on
15:42
that's for the foundation board we've been much less focused on sort of what happens with the money and how the money
15:48
gets reported after it passes through us um just just something to me just
15:54
something I wanted to be upfront about I should add that uh adding
16:00
it's it's quite possible that this could end up in the form 990 in the way you describe as an in-kind contribution a
16:08
grant uh effectively a grant to an external organization Maybe
16:13
but if we described that as profitability for the
16:21
Wikimedia Enterprise project I believe that would be understood as
16:27
trying to cheat because they are not giving us any money it is free so we are giving them
16:36
something of value but
16:41
trying to describe giving away something to us for free as improving our
16:48
profitability would be maybe that would work in an
16:54
accounting system but I think the community
16:59
would understand that as cheating because we are not actually getting any
17:05
money from it and so it does not help our profitability if anything it is an
17:10
expense because we have to provide something for free well that yeah just
17:15
to I want to give you a chance to respond I'll go but just to to hot pile on to
17:21
that comment I think um that is how we think about it but since I mean when
17:27
we're doing Financial estimation for the project we are actually making the assumption that there's going to be some
17:33
amount of money that we need to set aside to support um free use of the service whether it's
17:41
because it's a um you know a sort of a sister organization like internet archive that
17:47
we want to support or uh as I suspect might be the case in this coming year there's you know potentially other other
17:54
organizations or types of Partnerships where we'll just we'll see value for any number of reasons some tied to
18:01
Enterprise some tied to larger Mission goals I suspect to to um to not charging
18:06
we we just we've been baking that into our cost assumptions in the same way that if you still have a financial
18:12
report you saw that there's a um and over at 50 a 15 overhead charge
18:18
that we pay back which is pretty significant actually to the foundation in terms of our budget so there's just a
18:24
sort of an assumption that 15 of the money that we bring in just you know isn't counted in the way that we would
18:30
normally count it and uh I just I'm mentioning that because I
18:35
think we just sort of take all of these into account and then look towards profitability as a number that goes
18:40
beyond that and we feel I mean I feel like uh for me personally getting to this
18:47
this sort of Break Even moment that we're currently at as quickly as possible was sort of a very important
18:52
goal which is why we move towards it very quickly um the the next goal that we want to
18:58
move towards and I I can't say how fast we'll move towards it yet but the next goal we want to move towards is actually you know paying paying back the initial
19:05
investment so not just being profitable or uh achieving this so achieving profitability month over month so that
19:11
we know that there's more money coming in than we're spending month over month and then after that what we want to do
19:16
is say okay how quickly can we also pay back the initial investment I don't know how long that's going to
19:23
take us but that is from a goal perspective very much what we'd like to do next and quick quick as quickly as possible
19:31
but it takes it makes assumptions that that is again all with the assumption that there are these additional costs
19:36
such as supporting internet archive just baked into what we're doing
19:44
oh you're mutedly sorry yes I am um I hope that answers the question in
19:49
sufficient detail hooker but you uh welcome to clarify or add anything
19:54
Dennis in particular if you feel there's something being Lost in Translation
20:00
um please speak up I guess you just um had it fine but
20:06
maybe hope you can let us know more I hope your question has been answered in detail to hold you
20:13
yes it was answer thank you yeah I I yeah this is an important aspect what
20:21
you mentioned with maybe it's it could seem like a way of cheating if gender
20:27
Prophet seems yeah if there is a transfer payment from the
20:33
immediate Foundation to give me the others yeah the time what point I think and so yeah at the moment I don't have a
20:40
problem with it but I think at least I would be interested I think
20:47
it would be great if you find a way to publish
20:53
an amount but also I can't understand I don't think in some place yeah I think that's
21:00
important to pay attention to privacy and also think about what ranking is needed what not but I think it would be
21:06
interesting to get a bit uh uh and
21:12
understanding what part of the Services of are used for free and what
21:20
are and how much is it used through a contract where
21:26
organizations pay to I think that would be interesting if you publish their kind of ratios or
21:33
maybe to understand it a bit more yeah yeah do you mean in a sense of the
21:39
proportion of use which is by paying companies paying
21:46
customers compared to the proportion of use by free
21:52
companies and yes yes this is what I mean
21:59
uh Chuck might be able to get into that but my my guess is
22:04
not yet because we are too young to have any Baseline data
22:15
Chuck did you have anything you could speak to on the question of proportion of free versus paid over time
22:25
um I don't have anything substantial to add to that I don't have that data in front
22:31
of me and it's not something we're currently tracking as far as I know yeah I would say
22:37
uh the proportion here's what I would say the proportion of use for paying customers is much much much much much
22:44
more significant than it is for the free customers um in part that's because how a free
22:51
customer can access how a free customer can access it is
22:56
limited relative to our paying customers who have significantly more access to the data it can you know so for example
23:02
if you are using our snapshot API which is if you're familiar with our service is kind of the equivalent the commercial
23:09
equivalent of the the free Wikimedia dumps API um the free version of the snapshot API
23:15
allows you you can download it as many times as you want but it only updates once every 30 days or so once a month so
23:23
the likelihood that you're going to download that frequently is low because it's just the same data until it
23:29
refreshes a month later it's on a slower Cadence than the free dumps which refresh every two weeks but for our
23:35
paying customers you know they can download that once a day and they can also download hourly diffs uh so that's
23:42
a like a much more significant load and since we measure everything Based on data egress data out
23:49
um you know it's a as you can imagine even just on that one API of significantly more use for our paying
23:55
customers uh and that's you know uh equally true of the for example we have a um our real-time API which is what we
24:03
also call our fire hose API because it's the one that again if you're familiar with our free API Universe it's the one
24:09
that's that's comparable to the event streams AP or event stream API
24:14
um and that it's uh sort of firing all of the edits and all the changes that are happening across our projects
24:20
um that is not accessible to free users but it is accessible to some of our paying customers
24:26
um and so that's again a place where there's significantly more data usage so I I would say we could on the one hand I
24:33
could say we could try and get those numbers and pull them together but if I'm being honest I don't think they'd be very helpful useful partly because we
24:40
haven't had free customers for that long um and partly because I just like anecdotally I can tell you the
24:46
difference is going to be huge I should clarify the um
24:51
there's two categories we don't call them both free
24:56
um but from the General concept of the word there are two categories of three there is the
25:01
trial user the account on the website that anyone can sign up for and get the
25:08
monthly dump that's fine there's no um you don't have to talk to anyone to do it and you can also download the via the
25:16
existing Wikipedia Foundation dumps then there is the sort of special case three
25:22
version so if this is the internet archive for example they get access to
25:27
the same thing that Google has which is
25:33
as much frequency as they want so that is not automatically available
25:40
to for free to anyone in the world because it's a commercial service primarily but
25:45
secondly basically no one has a use case for that
25:51
scale the internet archive is a very rare organization because it has
25:57
a genuine need for that scale I cannot speak to
26:03
how they are using it uh yet or in the future that's their own internal systems
26:10
but there is basically no one who is a
26:16
non-profit organization or a volunteer or an individual even academic
26:21
who has a need for the right now everything
26:28
every change speed and volume commercial organizations like search engines do
26:35
need that do want that and they are paying for it because that's an expensive and difficult service to
26:41
provide and that is where all of the volume of data the volume of
26:48
commercial value comes from so the proportion of free
26:53
users um the the amount of data that is coming from free is very small not because it's
27:01
restricted but because there's no one who has that need for the
27:07
huge amounts of data that uh large commercial organizations require
27:15
since we are speaking uh on the the technical side
27:21
data volumes uh I hope that we can come back to you if you had a follow-up question on this
27:27
uh but for the financial perspective but I wanted to raise a question that was asked on the mailing list uh uh with
27:36
regards to the statement in the uh software update
27:43
saying I'm curious about the implications for optimization that was
27:49
mentioned in the the changes that have occurred over this last year and the word
27:55
optimization we have optimized that was written in the product update
28:03
perhaps Ricardo or chuck could speak to
28:08
what was meant by optimizing
28:14
yeah I can go ahead just a note before I go not that I would try to hide anything
28:19
I just wanted to give a note that I'm three months here so if I miss something I'm happy if you give any questions to
28:27
Liam post for it I'm happy to clarify them in margita just because this is going to be a record I wanted to be as
28:33
much as straight as I can so the terms of optimization is simple just like any software when you build the first
28:39
version We and the identify points that can be improved either points that can
28:45
give us Financial uh gain or optimization or give us performance
28:51
improvements from what we previously had so in that performance in that in that
28:57
regard what we try to do with this new releases is coming out is exactly that is up to optimizing internals of the
29:04
software in order to achieve better performance and therefore lose use less resources for example and other things
29:12
like optimization of the the code itself and making it clean and more
29:18
more Enterprise let's call it like that I hope that clarifies if somebody has
29:23
any other further questions I'm happy to clarify also
29:28
yeah and um in the product updates that's specifically what I was
29:34
um overshadowing to keep it less technical
29:42
um uh with the the blog post that we'll have on our news
29:47
page um I can get into a little more specifics about that but essentially
29:52
that's a it's like a code audit a restructure
29:57
um based on you know what we've learned in the last couple years of having the software written from a very V1 and a
30:04
beta standpoint to being in production how people are using it and how people would like
30:10
to use it in the future and features that we would like to build um so a lot of the
30:16
restructuring and rewriting of the kind of the architectural pace of that code um is helping with what we're gonna you
30:24
know have in the pipeline to build we needed this to exist to in order to build on top of that if
30:31
still trying to keep it a little non-technical there
30:38
so uh we'll make sure that uh the person who asked that question will get a link
30:43
to this particular answer for the detail uh Martin uh as the the other person in the
30:52
room who hasn't had a chance to ask a question specifically I wanted to particularly call on you if you did want to ask
30:59
anything or you you don't have to but you're here so I want to make sure you had a chance uh okay chat message you're
31:06
fine just listening great uh uh to come back to you
31:13
that did the previous answers or the previous things
31:18
are they clarified sufficiently answered sufficiently for your
31:24
um for your needs or yes you have other questions that you had not asked yet which we want to to go
31:31
to yes I have a question and I've then stroke me so I I asked it in German and
31:40
then he will translate and I want to try it then better yes I am not so good
31:48
in English I saw my name is
32:07
wikidata in the RP services to integrine or from Wiki data updates to the common
32:14
order give this foreign
32:20
foreign yeah and the question is are there any plans
32:25
um and if there are then when um to offer the same services with
32:31
Wikimedia Enterprises for Wiki data especially as far as I understand it's
32:36
part of the bigger thing already but maybe you might clarify yeah you want me to take this one please
32:44
okay um we yes we would very uh it has always been part of the plan to integrate
32:52
wikidata into the Wikimedia Enterprise product in some way the um the qid
32:59
structure already exists in our data set so uh was that was you know done
33:06
uh or we we use that as sort of the key identifier in our data set specifically with an eye towards you know a world in
33:13
which we could integrate Wiki data into our apis um we have been in constant
33:20
communication with the wikidata team uh at Wikimedia Deutschland because uh for
33:26
us to do this we have to ensure that we do it in a way that is uh as because
33:31
they are the team that supports it um it's very important to us obviously and also to them that we we integ that
33:39
if and when ideally when we integrate it that we do it in a way that is respectful of what they have already
33:46
done and is um uh is going to con I would say going to
33:51
contribute meaningfully to their goals um you know uh so a lot of our
33:59
conversations are really around that and working towards alignment on that and that's taken a little while but I feel
34:06
like we've had very good conversations with our team and we're definitely getting closer
34:11
um all of which is to say we don't have a timeline I think everyone on our team would love to see it happen this year in
34:19
calendar year 2023. um we know it's very important to
34:24
uh every customer that we've had or every potential customer that we've talked to
34:31
particularly for example like smaller search engines that maybe don't have the
34:36
um the resources available on their end to do the elaborate parsing work that is sometimes required to use not sometimes
34:43
that is required to use Wikipedia data for example um it's much easier for them for obvious
34:50
reasons as the knowledge graph to get started with wikidata so they would love to see it integrated larger companies
34:57
that we work with such as Google make extensive use of wikidata and would love to see it integrated as well
35:03
um we think it would be valuable for the product I think it could potentially be really cool so there's that due
35:10
um it's just a question of you know what is how and this is very much on my mind which
35:17
is why I'm making this space I want to I just I want to make sure that we find a way to do it that it's like uh
35:24
yeah that is like an agreement between us and we committed and then we could add a team where everybody is clear on
35:30
what uh clear on what's Happening and that everybody's getting getting sort of what they need out of the partnership
35:37
and that's that's what we've been working towards for the last year and I think we're
35:42
I feel good but I would not say 100 certain that we
35:47
will get it done this year and the first version of that integration in 2023.
35:53
um but I'm I'm not at a place yet where we can make any promises on that front because we're still in discussions with
36:00
them about what it will take to make it happen
36:05
I guess I guess what I'm saying is it's there's technical challenges but the things that we're trying to work through
36:11
that are bigger than that are not necessarily technical challenges they're just organizational and structural
36:17
challenges to make sure that it's um done right done in the right way
36:22
carefully and thoughtfully and in everyone's best interests
36:31
yeah thank you foreign I have another short question
36:38
how high was the down time in the last year so how
36:44
how many hours was the Wikimedia Enterprise API not
36:52
available last year
36:57
I feel like the best people to answer that question aren't here but maybe Ricardo knows
37:10
I'm sorry no oh I think the the answer is I don't know
37:16
the number specific number but we have a uh yes I put a link in the in the chat
37:22
to the status page status dot enterprise.wikimedia.com that is the
37:28
uptime statistics live which covers the last
37:34
90 days nine zero days which is 100 percent
37:40
we have a contractual obligation for 99.9 uh
37:48
which is different it says SLA not a SLO so this is a contractual
37:54
requirement as part of a legal and financial responsibility this is different to the
38:04
normal Wikimedia Foundation uptime uh what you might have recently seen that Wiki data or the Wikimedia Foundation is
38:12
talking about SLO a which is a best effort requirement in fact I'm not sure
38:18
if that's the technical description but it's the it is not a legal requirement but a we will do our best to meet this
38:25
standard we in the Enterprise team have a contractual requirement that has
38:31
Financial penalties if we do not meet that requirement this also includes
38:37
Amy's team which is did you answer the phone when I called you fast enough did you answer my email
38:44
asking for a reset password fast enough this is quite different to the way the
38:50
rest of the Wikimedia movement operates and because of that financial and legal requirement is much more expensive to
38:57
operate it requires a level of redundancy that is unnecessary for
39:04
normal humans but because we have the contract that
39:10
requires this we have to have a level of Technical and human redundancy in that
39:16
support which would be basically it would be a waste to use donor donor
39:22
money for that level of requirement
39:27
but if they want to pay for it fine it's also a reason why we are
39:33
hosting at least for the time being while we get from zero to up to speed why the hosting is on AWS is
39:42
on the external systems because the they are much larger than Wikipedia is and
39:49
they can be responsible for building the infrastructure that allows that degree of service requirement
39:57
when we are more stable we have all the customers that we we know
40:02
from this month to next month it's not going to be twice as much service requirement it will be much more stable
40:09
then it is much more um sensible to talk about moving that kind of Hosting inside the Wikipedia
40:16
Foundation but it would be an extreme Financial Risk to try and
40:24
put that legal burden on our internal infrastructure
40:30
which is not designed for it it can do it but it should but our the Wikimedia
40:37
foundation's internal infrastructure should not be held responsible to the arbitrary
40:43
and extreme legal and financial threats of the contract
40:49
uh just because that's what an external organization feels like it needs
40:56
the short answer is we are very good but that costs a lot of money to do it and that's why they pay a lot of money
41:02
to do it I hope that answers the question
41:09
uh Amy wanted to do you want to say that or you're going to read that out to a chat message
41:19
yeah sorry as if the chat doesn't translate through to the recording then sure you can read it out precisely the
41:26
recording doesn't doesn't kick uh so to clarify those uptime and customer support response times the guarantees
41:32
are for the paying customers we do not offer that kind of contractual guarantee to a free customer because they're not
41:39
paying for it basically uh so we can't give them a discount next month because
41:44
it's already free um there was a question in the chat from
41:51
Andrew uh well a hand up in the chat from entry would you like to
41:57
speak your your question yeah um so I hello everyone uh sorry about
42:03
joining a little bit late um I was looking through the
42:09
um financial performance portion and I saw that the expenses have grown to
42:17
above revenues and I'm curious uh what you project the
42:24
increase in Revenue to because there's only so many possible organizations that
42:30
could use this type of service so I'm curious how you anticipate your growth to look say over the rest of this year
42:38
or later down the line yeah I can take that Liam if you want
42:44
um we feel so okay so depending on how you look at
42:50
the market that we're serving and the way that I'll just give you sort of my take on it the way that we look at the
42:56
market is there's a you know broadly a very very wide range of uses for
43:01
Wikipedia content Wikipedia data at Large across all of the projects
43:08
um and you know in some of the research that we've done and poked around it's actually kind of amazing how widely it's
43:13
used I mean I guess it's not really it's not surprising but it still continues to amaze me uh that said you know you can't
43:20
start a product and say oh we're going to sell to everybody because then you don't really have much Focus so our
43:25
focus on our initial Market as we describe it you can this is I think fairly clear from looking at the homepage of the website is search
43:32
engines and voice assistance um while it is true that that's not a huge
43:38
Market uh some of the players in it have a lot of money as you know
43:45
um and actually there has been kind of a Resurgence in the last
43:50
year or two uh and folks that are actually going after the search engine Market again that was true even before
43:56
the kind of the recent chat GPT if you're familiar with what's going on with um openai and the AI space there's
44:03
also a renewed interest in Alternative forms of searching through artificial intelligence and that actually also kind
44:10
of fits into when we look at the customer profile for what we think of as search engines and voice assistance
44:17
uh folks who are looking at uh replacing not necessarily search but searching
44:22
with artificial intelligence very much fit that profile as well so it is actually even compared to when we
44:28
started thinking about this as our Market two years ago it is actually a grow it is a renewed and growing Market
44:35
um and one that's fairly well funded so all of it just to say like that's kind
44:41
of why we are focused on this market and I think we'll continue to focus on this market so there's there's other markets
44:47
that have expressed interest uh that I don't think our product is really well suited for at the moment education is
44:54
one there's a lot of interest in um uh with a lot of interest from
45:00
educational companies they want something that's a little easier to make to put into like a question and answer format
45:06
our apis and our structure currently does not allow that but I can see a future where that would be it would be
45:11
possible to tailor it more towards something that looked like a question and answer format
45:17
um Financial companies have expressed a lot of interest uh separate apart from the search capacity
45:23
for artificial intelligence there's a lot of use of um Wikimedia data for training purposes and while right now
45:30
most of that's being done with the free dumps I think there's potentially Market opportunity there as well although for
45:38
all the reasons you can imagine we want to tread carefully when when thinking about that as a market
45:44
um all of it is to say I actually think the potential from a business perspective is quite High
45:49
uh in each of those markets we have to figure out what they want to buy you know which is made you know part part of
45:56
the child part of the challenge for us always is that uh we're trying to sell something that we also give away for
46:02
free and that's frankly quite hard so you have to think really and unlike most com like if we were a more traditional
46:08
data sales company we would just not give it away for free and everybody would have to pay and make a lot of
46:14
money and go home obviously that's quite counter to the mission of
46:19
um the movement so that's not what we do instead we have to come up with more nuanced ways to figure out how to build
46:25
alternative approaches or alternative services that will have value in these markets
46:30
um I would say I don't I don't think it's a small Market I think it's actually a growing one I would agree that's overall I would agree that the
46:37
search engine and voice system Market uh isn't huge but we've really only begun to sell into it and I think we have a
46:44
lot of prospects for this year amongst both um some of the large Global search engines that we have not yet sold to and
46:51
also some of the new intros that the search Market that are starting to show a lot of promise and uh and have a lot
46:57
of capital to do it um and that's where a lot of our Focus they share is going to be
47:02
uh I don't have projections for this year quite yet in terms so I can't share
47:08
those out and uh any projections that we come up with actually have to be um approved by the board before they're
47:15
actually by the Wikimedia Foundation board before they're actually made public so that will be something that
47:20
will happen in the coming months as part of the annual planning process that's kicking off right now
47:26
um but I can say that I feel very confident and I think the whole team feels very confident that we will be able to get into a place of comfortable
47:32
uh month month over month profitability uh in Cal in this in calendar year 2023
47:39
um we have some really good sales prospects in front of us and a lot of a lot of I mean just a lot of interest in
47:45
being able to make use of this product in a more with the commercial guarantees that Amy was talking about before that's
47:51
actually a huge problem for a lot of the larger search engines in particular and so we see a lot of interest we think we
47:58
can do it the fact that we're already a break even is the reason that I feel pretty confident saying and the fact
48:03
that we're both at break even and not planning on expanding the team too substantially Beyond where it is right now suggests to me that we should be
48:10
able to get into profitability again month over month profit ability fairly quickly
48:16
with the emphasis on the Fair Lake because I don't know how long exactly it's going to take and don't want to make any promises
48:24
helpful thank you okay yeah I'm always happy to discuss this in more detail so
48:29
any of that needs clarification let me know I did have a bit of a follow-up um and
48:35
you slightly touched on that in that uh Wikipedia and Wikimedia products in
48:41
general have been from the start uh marketed as
48:46
free accessible to everyone everyone has access to all the same information in
48:54
theory eventually uh I mean already more information than anyone can
49:01
ever consume in one lifetime um and so I'm curious how
49:08
you reconcile this notion of selling a um
49:15
of a not-for-profit uh governing entity selling this data to for-profit private
49:25
corporations and what that would look like fur an average
49:32
human trying to use a Wikimedia product as it was originally
49:39
described certainly I this is a question we obviously deal with as a fundamental an
49:47
ideological level not just at a commercial advertising business level
49:53
but it's this project is part of the Wikimedia Foundation part of the wiki media
49:58
movement so it has to operate within not against that those principles
50:05
um you know as a long-standing Wikimedia volunteer myself that's that's why I'm here
50:11
as a wikimedian not not because I'm trying to uh uh run counter to it but the the two
50:20
reasons the two issues is wicked media content
50:27
has always whilst the organization is non-profit and non-commercial websites
50:32
don't have ads and you know all that jazz which you you and I know
50:38
um has always been available for anyone to use for any purpose including commercially
50:43
and these organizations to whom this project is selling the API access are
50:52
already using Wikimedia content extensively and making profit from it
51:00
their use of that information costs the movement money and time and effort to
51:07
maintain the service to them it is not free because of the the high
51:13
requirements they place on the infrastructure it's the method methodology we sorry the
51:20
metaphor we use is imagine if a large Factory attached itself to the city
51:28
water supply and then said you need to give me industrial quantities of water
51:35
at the same at high pressure that you that you give also to the
51:42
individual people the individual houses that cost the city a lot of money to
51:47
provide that water even though the water is exactly the same so if the city gives it to those
51:54
companies you at the same price or the free like it
52:00
gives to the rest of the the community that means that that the communities
52:06
uh the service provided to the rest of the community is diminished because they
52:12
have to concentrate on the largest loudest heaviest user of the service
52:18
so we this project for answer one is this project flips that requirement they
52:26
pay us and subsidize the Wikimedia movement provide Revenue diversification
52:32
to the Wikimedia movement for their use of the same information at
52:39
high speed instead of what has happened all the time until this last year which is the
52:44
Wikimedia movement and donor money is subsidizing them and their requirements
52:49
and I think that's unfair um it's the classic tragedy of the
52:55
commons story or here's here's an available field one large farmer comes
53:00
on and and with all his cows eats all the grass uh okay if he wants to have
53:05
that many cows we can have a separate feel just for that farmer uh and
53:11
everyone else can have access to the normal field the second reason is
53:16
this is a crucial distinction this project is not selling data it is
53:22
often it is easy to misinterpret sometimes misinterpret deliberately by
53:28
people trying to um find a controversy that does not exist
53:35
but most of the time it is misinterpreting because it's standard that an API or an API company sells the
53:42
content if you are a weather API if you are a traffic API you are selling the
53:49
traffic information or the weather information well the finance information we are not selling the information we
53:56
are selling the pipe the information is the same and it's freely licensed in the
54:02
case of wikidata it's cc0 so it does not even require attribution by law
54:08
the content is the same we structure it a little differently in terms of the API
54:14
metadata but none of the structure or none of the information in that metadata is new or
54:21
secret it's things like does this page have a different Sudden
54:28
Change in um readers number of readers that's something that wikimedians already use or does this page have a
54:36
sudden number of uh IP editors Anonymous editors that's information that wikimedians already use to make
54:44
decisions about should the article be locked for only administrators or should
54:49
we um does it need temporary protections
54:54
that kind of information has not been available in an API before and we're trying to add that in but it's it's
55:01
already existing data so what we are selling is not the water
55:06
to return to my original metaphor is not the water we're selling the pipe and the
55:13
contractual guarantee that the water will be provided at a certain
55:19
water pressure and that there will be a telephone available for you to answer uh if you
55:25
have a question it's quite different from uh the idea of selling better water
55:31
which we do not do that water is already available for everyone for free
55:36
I hope that those two issues responds to the question of the sort of
55:42
the the ethics of what we're doing Andrew uh that does provide some clarification
55:50
yeah um and then also uh I'm curious uh there have been a number of instances where uh
55:57
especially Google uh to the extent that there's even a Wikipedia page called uh
56:06
Wikipedia are called Google and Wikipedia um where Google has donated large sums
56:14
of money to the Wikimedia Foundation over the last multiple years and so I'm
56:20
curious uh how many other companies to which you're
56:26
marketing this project already donate to the Wikimedia movement in some capacity
56:34
and do you anticipate those donations would go down somewhat proportionally to
56:41
the amount that you're going to be charging them for API access
56:47
one of the yeah Lane did you want to answer that or oh uh well I'll just say I mean I'll just I would like you to
56:53
answer it actually but I'll just say this is actually a was a topic at the very beginning of the when Enterprise
56:59
was first getting going this was actually uh the topic of whether or not companies would would see you know their
57:06
donations versus sort of commercial payments as um uh you know complementary or in Conflict
57:13
where one would go down and the other would go up and there wasn't a lot of certainty and in fact in those cases
57:19
where the customers have been had donated it's it seems like or or uh or
57:25
considering it it seems like it's very it's very Case by case some of them treat them as very separate because of
57:30
the way their organizations work some of them absolutely see the money moving over from one to the other
57:36
um so it's not it's not it hasn't been some of them haven't ever donated and it's just not
57:41
the way that they're because they're not large enough um or because they're not the kind of organization that donates money no
57:46
matter how large they get uh what what I what I will say though is
57:52
that um you know in this one of the reasons that Enterprise and the idea of
57:57
Revenue diversification around Enterprise with corporations was a priority uh from the financial
58:05
side of things was because it is it is a more consistent and stable guarantee of income you know every year that a
58:12
company donates they have to make the decision to donate every single time and you know our goal is to sign multi-year
58:18
contracts where we're not necessarily locking them into our service but providing them with additional value in
58:24
such a way that renewal of those contracts when the time comes to renew is an obvious thing for them to do that's kind of how we think about our
58:32
business and what we're laser focused on doing and that's just a lot more comforting I think when it
58:38
comes to working with commercial organizations than kind of going hat in hand some year after year and hoping that they'll want to make the same big
58:44
donation next year that they did this year particularly in this moment where a lot of them are pulling back from a lot
58:50
of their expenses and even before they let go of 10 000 people on their staff they let go of giving away any of their
58:57
money for philanthropic causes so so yes yes and sometimes yes sometimes
59:03
no sometimes and overall a hard Revenue stream to depend on in any sort of
59:10
consistent or planful way when uh when it's donations
59:16
do you have anything to add um I think pretty much covered it
59:22
there's there's an issue of how uh those kinds of companies deal with donations versus contracts as from their
59:30
perspective tax um but primarily as as Lane said
59:36
it is a more consistent I mean these kinds of companies Google and so forth have various kinds of relationships
59:45
um with Wikimedia movement funding Affiliates and sponsoring conferences
59:51
sometimes and so forth and this is trying to be deliberately not owning the
59:57
entire relationship with the rest of the organ of the movement uh we're just talking about API access
1:00:05
but in as much as that might affect that company's interest in donating
1:00:11
I think the Wikimedia movement is much more
1:00:16
that is much better served having a relationship of
1:00:23
of a a formal legal relationship like this does with a giant company
1:00:28
than going to them every year every two years begging for a donation because it's
1:00:36
quite different principle than donating the movements focus on donation or the community
1:00:42
foundations focus on donations from small donors to the five dollars the ten
1:00:48
dollars that's crucial for our independence as a movement not having a
1:00:54
large proportion of the money coming from one big donor who then gets to
1:00:59
influence the organization politically and we have checks and balances balances on the maximum amount of any individual
1:01:07
company or total we can obtain as well but having a contractual clear
1:01:15
relationship that is long-term with these kinds of companies is much more
1:01:20
stable financially and much more there's more confidence in our
1:01:25
independence and what we owe and what they owe us then writing once every year or two years say
1:01:32
hey please could you uh donate some money because you know you use this a lot uh that's
1:01:40
they are making so much money off the Wikimedia knowledge that we should stand
1:01:46
up and actually say that they need to invest in our movement rather than
1:01:53
be look magnanimous every time they happen to donate a little bit here or
1:02:00
there is a question in the chat from Martin I
1:02:05
wanted to not forget um Ricardo could you speak to this the question is
1:02:11
um the additional features we mentioned earlier uh in the in the update about
1:02:17
how we're building out the software are these being built due to customer
1:02:22
request or is there a strategic plan for developments to the API
1:02:28
uh I I assume this is not an either or but a both yeah I think is yeah go ahead Lane no no
1:02:36
you go please yeah okay yeah it's a it's a big a mix of
1:02:42
of both um as a revenue a revenue generation company let's say like that we listen to
1:02:50
our customers and their request as features we have a clear example for example the summary delete section of
1:02:56
web page is offered by the Wikimedia apis giving a summary and the customers
1:03:02
requested us for example to give a different perspective on that that we
1:03:07
are start offering but as we are offering and releasing that at the same time we're giving back to the Wikimedia
1:03:14
and there there are conversations going on to the development we did will be put
1:03:20
back in the public API so that we continue to be using Liam's terms we continue to be the pipe and not any
1:03:26
content generation so that we don't allow companies to influence that content in any way or Etc that's my
1:03:33
perspective and uh that's how I see it so in terms of that is there is a
1:03:38
strategic plan we have product managers clearly designing where we want to go and how are you going to go but we
1:03:44
listen to our customers and to the companies we interact with in order to go after the revenue
1:03:56
tendency the question you had with regards to planet yes I see in the chat uh you're happy with with that answer
1:04:03
there is a roadmap listed and a quarterly update listed on the media
1:04:09
wiki page for Enterprise which is obviously technically focused and Chuck
1:04:15
will be publishing on the Wikipedia Enterprise news page uh in the next couple of weeks a more
1:04:22
technically oriented update about here are the new features available that's not describing the roadmap or the
1:04:29
Strategic plan in general but describing the features of now that are coming in the next week or two
1:04:37
um but on our media wiki page in general is a roadmap Uh current development
1:04:42
priorities Etc including down to things like the fabricated Board of what is the
1:04:49
individual bugs and things being worked on this week
1:04:55
depending on your level of strategic thinking about feature development
1:05:04
uh were there further questions from anyone because that is the list of questions that were written in the chat
1:05:12
to now Martin follow-up question what if a customer request exceeds the time of
1:05:17
your developers do you hire more developers if it makes sense to fulfill this request can you use the resources
1:05:23
of the Wikipedia foundations teams itself good practical question it is yeah I
1:05:30
mean we don't do every we don't do everything they ask for um and it and it tends to be more of a
1:05:35
sort of a trying trying to understand uh particularly of life some of the work that we're trying to do
1:05:42
in um making the uh as you might have gotten from the product update some of
1:05:48
our Focus right now including like for example integrating Wiki data is under what we call the heading of you know
1:05:53
machine readability which means you know both like wikidata which is obviously designed to be machine readable so
1:06:00
integrating that but also looking at looking at the um looking at the the
1:06:05
what we can extract from the data say for example a Wikipedia page or
1:06:10
Wikipedia article um like what we can extract from that that would be useful right so not trying
1:06:16
to make it all machine readable all at once but saying okay we know certain parts of this page are useful and
1:06:21
necessary and what can we pull out of there and so one of the the product manager Stephanie is unfortunately not
1:06:27
here who's been working on that but part of the way that she's gone about it is she's just talked to a lot of current customers and potential customers
1:06:34
um that are very interested in that feature to try and understand okay what are the parts of the page that would be
1:06:40
most useful for you to have um so it's so it's really just more a
1:06:46
question of like okay starting to understand that allows us to prioritize um better but I don't know that we're
1:06:53
necessarily going to build them everything that they wanted uh uh so that's sort of how we do it I
1:06:58
would say we we are not looking to hire additional or not we are actually in the process of hiring some more developers
1:07:04
but beyond that that's sort of to build out the team for what we see in front of us
1:07:09
um but uh but we're not looking to sort of scale up significantly beyond that in terms of the engineering team rather
1:07:15
we're trying to make sure that the scope of work that we're doing is scale scaled to the number of Engineers that we have
1:07:21
um and as for working with resources from the foundation itself we generally try to shy away from that the overhead
1:07:28
charge that you may have that you may have seen in the financial report that I was discussing earlier it's not really about pulling in engineering or product
1:07:35
support it's more um uh sort of more business oriented support so it's about paying for the
1:07:41
legal team it's about paying the legal team's time it's about paying for the time of the folks who do RIT Services
1:07:47
it's about paying for Finance and Accounting codes um
1:07:52
well I think we're generally speaking we're trying this I mean there's a lot of work to do on the foundation side to
1:07:59
support the technology over there and our goal is to um to our our goal on our team is to
1:08:05
pull over as much of the overtime as to pull over as much of the commercial activity on the our apis as possible
1:08:11
specifically so that it's less work for the folks over at the foundation to support commercial usage in the way that
1:08:18
Liam was describing earlier so so generally speaking we would prefer not to use the resources of the rest of the
1:08:23
foundation whenever possible there are certainly times when we talk to them about things that we need or aspects of
1:08:29
like parts of the foundation's core infrastructure services that we would like to see improved but we sort of see
1:08:35
ourselves as just one one voice and one stakeholder among many when those teams are making their decisions
1:08:41
we also don't get to keep our own money in the sense oh yeah the
1:08:47
money passes right there it's by yeah it goes to where the rest of the money the donations and so forth go to and then we
1:08:54
ask for resources like every other team and that's appropriate it means you don't get everything you want but it
1:09:00
means just like the fundraising team doesn't get everything it wants even though it was the one that got the money
1:09:06
in the first place uh um Amy I see your hand up to answer this question there's also a question in the
1:09:11
chat from Andrew I will get to that next thanks yeah I was just gonna just um
1:09:16
further clarify on that little point just um you know our customer it's not client services so you know our
1:09:23
customers they're not dictating a timeline for us on things that we're doing you know they might have requests
1:09:29
and we have conversations with them about it but of course it you know has to fit in with our our bigger strategy
1:09:35
as well and and we you know figure out what timeline works for the scope
1:09:41
um and size of our team and our own resources and um you know so you know
1:09:46
there's the question about like if uh if their request exceeds
1:09:52
um sort of our ability do we hires and you know so yeah we we don't do that and
1:09:57
they don't they don't dictate to us like when something has to be done by or you know it's different in that way like
1:10:05
what what a client services sort of a model might look like
1:10:14
okay thanks Amy uh Martin seems in the chat happy with with the detail of the response there's a question here from
1:10:20
Andrew would you like to read it yourself or should I read it out for you
1:10:25
yeah I was just curious um if the company
1:10:31
comes to the uh Enterprise team and asks for a specific API feature to be made
1:10:38
available to them would the people who spend hours every day making Wikimedia
1:10:48
what it is would they have a way to potentially object to the
1:10:55
creation of such features and will such objections be taken into consideration
1:11:01
by the team yeah I love this question um it's really it's interesting I mean
1:11:08
you know the way that we're trying to approach the product roadmap that we put together is by sort of you know again
1:11:15
sort of doing the research understanding what it is that companies are asking us for figuring out you know
1:11:20
uh where is what like the range of companies that we're trying to serve right now are trying to pull aboard as
1:11:25
customers you know where where is there the most overlap right where can we get the most most for our Resources by
1:11:31
focusing on like a particular feature that we know the vast majority of them want that's the ideal um and then we roll all of that up into
1:11:38
a product roadmap um which we do publish publicly and maybe you can drop the link in thank you
1:11:45
right there I'm meta oh that's the principles um but yeah there's the so so I kind of
1:11:51
two answers to this question one is like we're trying to publish it as a product roadmap and I think that if there's a
1:11:56
desire to sort of intervene at that point in any of the features I think that would be uh thank you uh I think
1:12:03
that would be a really good place to intervene um uh and then and then there's this more
1:12:09
abstract layer of the principles that we put together and I really I have to commend Liam for this he really in a
1:12:16
similar way that I was describing really did the leg work to understand early on in the first years like what the
1:12:22
community uh Community ideas or Community concerns were and what it what what we needed to
1:12:27
get into the principles both in terms of sort of positive or aspirational goals as well as you know defensive promises
1:12:34
things that we said that we would never do um so I would say in either of those
1:12:40
cases like we're perfectly happy to continue to have a discussion the reason I said I love this question it's like the team would uh be overjoyed to
1:12:46
continue a discussion about the principles I think like any document it should be a living documentum I don't want to remove principles that are
1:12:54
keeping us from uh or that are keeping us holding us to our commitments uh I
1:12:59
still think there's plenty plenty of room for a conversation around those to see if there are things that need to get added or modified and meaningful ways so
1:13:05
that's one level and then I would say on the product roadmap that we we are trying to publish that
1:13:12
um I I had a head of actual product development on a fairly regular basis
1:13:17
um sometimes we do a better job or Worse job on that but we're trying very hard as we figure out what it is that we want
1:13:23
to build and I think would be a great place to engage with us um I think we could also potentially
1:13:28
discuss we tend to do these public office hours around kind of announcements and events but that could
1:13:34
be another place where we could potentially start doing it in terms of product or feature discussions if that was something that folks in the
1:13:40
community would be interested in yeah I meant oh go ahead
1:13:46
so I was I was going to speak uh two exactly the same two levels that you
1:13:51
were referencing there so there is so under the broadheading of community
1:13:57
oversight of features or Community oversight and and interrogation or second opinion about software
1:14:05
I would say it's the same as anything else in that the Wikimedia Foundation product or technology departments build
1:14:13
there are various informal methods of
1:14:19
of raising concerns or expressing opinions on
1:14:24
high level areas like an annual plan down to an individual commit or an
1:14:30
individual feature because we have the fabricator board and you can comment on a fabricated ticket so those kinds of
1:14:36
things are all the same as any other given feature development
1:14:42
process for any other Wikimedia Foundation product um tool there's also the formal
1:14:50
methodology things like board you know appealing to the elected members on the
1:14:56
Wikimedia Foundation Board of Trustees which is your kind of official formal legal method of of
1:15:03
raising those kinds of concerns uh and you'll be unsurprised to know that I
1:15:08
spent a lot of my time answering questions on the top page on meta asking
1:15:13
similar questions to the the ones we're discussing today in the chat uh hogu can
1:15:21
perhaps confirm that I'm fairly fast in answering questions and we can try and
1:15:28
get back quickly precisely because we know that this is a weird and unusual
1:15:34
thing inside the Wikimedia movement I would like to think that we are trying
1:15:40
to be scrupulously available and clear about what is happening
1:15:46
not that anyone else is being unclear deliberately but because there is more
1:15:52
scrutiny on the potential ways of
1:15:57
misuse of what this project is doing that our road maps our plans our
1:16:05
adherence to the values of the movement are more uh scrutinized then a lot of corners of
1:16:14
the Wikimedia movement the so that's one and two is the
1:16:20
principles that I link to the chat there which is if you go to the Wikimedia Enterprise page on meta and there is a
1:16:27
sub page on the link from the info box principles that refers to things like these are the guard rails we've set
1:16:33
ourselves and they're not in law but they're written there
1:16:39
be loud and large so anyone can hold us to account for it uh things like
1:16:46
when we try to write this in a way that's future proof so things like there is no exclusive contracts no
1:16:54
exclusive content so we're not building something for One customer
1:17:00
and then they get to exclude other potential customers their competition
1:17:05
whatever we build is available for everyone including the free users of the
1:17:11
trial service or uh non-profit organizations who wish to have the the
1:17:16
free version as well there is no exclusive content so we're not building
1:17:24
some getting some secret data feed
1:17:29
and selling that data feed so that the large commercial organizations get a better thing
1:17:36
there's no um special information within the data feed that is not already
1:17:41
available perhaps not as easily findable but not secret or new
1:17:47
those there's a bunch of other principles in there about financial transparency and so forth but I think
1:17:53
those two things are quite crucial for ensuring future proofing
1:18:00
the appropriateness of the features built by
1:18:05
the Wikimedia Enterprise team and of course we're always working with the with and under
1:18:12
the same privacy policy in terms of use policy Board of Trustees over a site
1:18:19
audit committee anyway so we can't make up
1:18:25
rules as we go along it's still part of the bookie video Foundation I hope that uh and and all the law that comes with
1:18:33
that I hope that answers your question that was helpful uh though to clarify
1:18:40
something that you'd said I wasn't necessarily objecting to uh or I wasn't
1:18:45
necessarily referring to a potential customer requesting exclusivity I was just
1:18:52
referring to the creation of or providing easy access to a feature in
1:19:02
general rather than to that specific company just like for example if
1:19:07
um I mean if you mentioned uh some
1:19:14
one example that you mentioned was um which pages had a high influx of Ip
1:19:21
editors if you made that list then someone could go to that list and see so
1:19:28
here are all the IPS that are editing this page and here are where all these
1:19:33
IPS are based and then make some potential inference about those IPS uh
1:19:39
which see there's a lot of security uh there's a lot of data that is available
1:19:45
that becomes a security it's public but it becomes a security or privacy risk when made available at high speed or
1:19:52
aggregated yeah uh even though the data itself is already there and so we're
1:19:58
very careful not to do that unsurprisingly uh and we also we are
1:20:04
only talking about the Articles here we're not talking about user Pages we're not selling or
1:20:12
including um editor information it's the content not
1:20:18
the the use of the users themselves the so we can be the foundation legal
1:20:24
team is very working with us a lot about the formal requirements for privacy
1:20:32
and uh reader protection but equally there they are
1:20:38
considerations that apply already to the existing apis and existing database
1:20:44
dumps because all of this content is already available through apis and large commercial organizations are already
1:20:52
scraping as much as they possibly can from Wikipedia
1:20:57
sites so we are trying to make it a little bit more structured a little bit more
1:21:04
cleaner and thereby bring back some degree of control or power to the Wikimedia
1:21:12
movement in how the data is used and how the data is expressed
1:21:17
rather than just saying hey anyone it's it's available technically it's true from a legal or
1:21:24
from a copyright perspective um but the considerations are a bit more
1:21:31
technically and legally structured in in an API rather than just scrape everything
1:21:37
we also I should just briefly mention that because of that non-exclusivity
1:21:43
principle I mentioned before a lot of people are
1:21:48
initially concerned that this kind of service will benefit the biggest players
1:21:54
and increase commercial monopolies on access to data whereas although they
1:22:01
might be the first customers we feel that because the data we're providing is
1:22:06
the same to everyone and the price just depends on how much you're using it
1:22:11
this will benefit these smaller players to help
1:22:16
get them to be able to use bigger media data in a way that they have not been able to do before this will level the
1:22:23
playing field rather than increase the power of Monopoly of larger
1:22:28
organizations so we think it has an equity aspect to this too not nearly for Revenue but
1:22:35
companies and organizations small search engines that have never been able to use Wikipedia data because it's just too
1:22:42
damn hard will now be able to do so increasing the reach of the content and
1:22:48
getting more people to see more things through more diverse methods of access not just through the
1:22:55
one big company okay that's a little bit
1:23:03
more reassuring but um I mean the the reason I brought up IEP users editing a
1:23:10
specific page was because when I joined that was an example that was mentioned in
1:23:17
response to my question so I mean that's
1:23:23
I don't know uh and then finally uh this isn't necessarily a question about
1:23:30
uh the Enterprise project itself but in general about the
1:23:36
structure of Wikimedia and how it communicates with its users and that is
1:23:42
the only reason I knew about Wikimedia Enterprise and by extension
1:23:48
this Zoom call is because I'm in a telegram
1:23:53
Group which notified me that there was a zoom call happening um though
1:24:00
uh the subsequent um Wikimedia uh meta web page for this link
1:24:09
said uh 9 pm UTC which I think is in half an hour so that slightly threw me
1:24:16
off as well and then the reason I'm in the telegram Group is because I was in
1:24:21
another telegram Group which said hey if you want to get all these updates about
1:24:27
the Wikimedia project then you should join this group and not the one that I was in previously and
1:24:35
I mean I can't imagine how many people are on my shoes I mean I've been on and
1:24:41
off more actively since early 2020 but on and off I've been a Wikipedia editor for
1:24:49
a little over a decade now and I feel like
1:24:54
so many initiatives whether it's this or whether it's some other feature
1:25:01
um are being rolled out with feedback from
1:25:06
at most I don't know a couple hundred distinct users whereas the number of
1:25:14
people who use the Wikimedia project in some capacity
1:25:21
or another is on a daily basis I would imagine at least in the hundreds of millions
1:25:29
um and so just this entire structure it's definitely not the fault of the
1:25:36
Enterprise project but I feel like that's something that
1:25:41
should be looked at at some point so I can I can speak to this but I
1:25:48
noticed uh Amy had her hand up from before so I want to check if you wanted to ah go back I'll be really yeah I'll
1:25:54
be really quick because I wanted to address Andrew's concern about the IP address thing because I think Andrew did
1:26:00
come in in the middle of a statement and I think lost some context um there
1:26:05
um missed some context rather um and I think I can address that quickly where I think the second issue
1:26:11
Liam definitely should handle it might be a much bigger response um um I can't remember exactly the example
1:26:17
Andrew but I think um it was more of maybe a hypothetical or laner Chuck maybe could clarify more
1:26:24
for me but the idea is just that there could be a um you know we're doing some work around like credit we're calling
1:26:29
kind of credibility or credibility signals or content Integrity but the idea would be that like maybe there's
1:26:35
something could get returned saying oh this article has suddenly had a lot of anonymous IP address edits on it
1:26:43
that might be something of concern that doesn't mean it's returning a list of what those IP addresses are just a
1:26:49
signal that there was a bunch of Anonymous IP addressed IP address edits
1:26:54
on it and that's that's that piece of um data is could might be significant it
1:27:00
might be something to make a decision about whether or not you want to trust that particular version of that article
1:27:06
due to that or if that article should be like shut down and only for admin editing at that moment or something it's
1:27:12
not about returning a list of all those IP addresses that is not a feature that we would be building or ever be giving
1:27:18
to a customer um chucker lane or anybody else want to
1:27:24
um say anything more about that please feel free but I hope that puts that any concern around that to rest
1:27:29
um Andrew because for sure for sure for sure we are not in the business of getting that kind of data yeah to
1:27:35
anybody yeah we don't even that's not even in the API data that we're flowing
1:27:40
through anyway um and that's what I was going to mention but you did it a little more eloquently than I could have but yeah
1:27:46
that's that's not a concern because we don't have that data to give
1:27:54
uh I didn't realize that's what my answer was potentially applying we were we were talking about doing
1:28:02
and the the question of comms Communications strategy in general is a
1:28:08
continually fraught one because you have this equal problem of wanting to tell and get as
1:28:14
many people to know about athene as possible but like any organization uh there's so many
1:28:21
different things going on simultaneously that people suffer from overload of
1:28:27
being informed or being asked to consult or communicate about
1:28:32
everything simultaneously uh often it is frequently the simultaneous
1:28:39
concern raised from uh individual Wikipedia volunteers saying I wasn't
1:28:46
informed about X and I'm being told about
1:28:52
and asked to join consultation calls or information things about so many things
1:28:58
that I can't read them all and those are both true but also contradictory so the
1:29:06
Wikipedia Foundation has gotten better at this over the last couple of years with a more
1:29:12
um what they're calling a air traffic control making sure that announcements
1:29:17
and requests for comments and things are spread out rather than all on the same day and also different
1:29:25
that there are certain groups who are of interest to different kinds of announcements maybe in the Indian
1:29:31
Community is something that this announcement is important for and then another one is for the
1:29:37
Wikimedia um developer Community which is a different but in for some people it's an
1:29:44
overlap because they're in both of those groups um and so just trying to be able to
1:29:49
communicate to the relevant people but not to irrelevant to people is a hard task to
1:29:57
achieve when you're trying to do be as accessible as possible without overloading people's attention or
1:30:04
pretending that they should really care about things that are not relevant to them well Community Enterprise is not
1:30:10
relevant to most people in the Wikimedia movement because it's a technical feature that
1:30:16
nearly anyone in the Wikimedia movement has no use for so we do not want to try and draw too
1:30:25
much attention to ourselves by broadcasting hey you know Wikipedia Enterprise has an announcement and
1:30:31
everyone should read it because most people don't care and it's not relevant and that's fine uh I put that message on the telegram
1:30:39
group at a couple of hours ago by way of a little bit of extra advertising uh on
1:30:45
the recently created announce telegram Channel but uh
1:30:52
things channels that have that are for announcements only notably the Neta
1:30:57
forum are notoriously bad at drawing with where I did put this report of this
1:31:05
event this meeting a notoriously bad at drawing uh the right attention because most announcements are
1:31:13
irrelevant to most people you just care about the one thing that you care about that announced channel on telegram was
1:31:21
created last month precisely because of a frustration in the general telegram
1:31:27
Channel but there were too many announcements and most people were annoyed by it being flooded with announcements so it was pushed to a side
1:31:33
Channel which people don't really follow so it's a it's a circular problem if you
1:31:40
can think of a place where this call and this project should be advertised
1:31:48
promote like raised awareness that it's not that where the audience would care
1:31:53
please tell me and I will promote things and raise awareness about this project there but I don't want to
1:32:00
over announce Wikimedia Enterprise to groups who who would just get annoyed if I was
1:32:07
talking about commercial apis to them too much
1:32:15
so in general I agree with virtually everything you said just now
1:32:20
um and uh for the the other telegram
1:32:25
group that I was in isn't even the General Wikipedia one I don't I didn't
1:32:31
know that there was a general Wikipedia telegram this was a group about
1:32:37
um consisting of people who wanted to
1:32:43
discuss a board selection process either in 2022
1:32:48
or maybe even the 2021 and that just has since
1:32:54
gone to include various other discussion topics
1:32:59
um but the um would the
1:33:06
with the question of announcements I think that um there are certain things
1:33:13
that should be announced to a limited audience but on the other
1:33:20
hand there are certain things that should be announced to
1:33:25
the general public on the
1:33:30
top of the Wikipedia article for example and
1:33:36
um every now and then
1:33:41
when I'm logged in the only things that I see at the top of an article that our
1:33:47
announcements are either um vote in such and such election after
1:33:56
candidates have already been announced
1:34:01
um or things like Wiki loves monuments for example which I agree I don't really
1:34:09
have much interest in that but um something like
1:34:16
I don't know where the Enterprise announcement would
1:34:23
fit in that because on one hand it is a really big change
1:34:30
and a really significant proposal but on the other hand like you said it's not something that most people would notice
1:34:38
um this announcement was made to a relatively limited group because it's a interim announcement it's a finance
1:34:44
update the announcement of the creation of this project which was a year ago uh
1:34:51
almost two years ago actually we had several rounds of much more wider discussion in the formation period in
1:35:00
the hey we want to do this what do you think period during the strategy process
1:35:06
a paid API or equivalent was mentioned in two of the eventual strategic
1:35:12
recommendations um so there were and there is a section on our on our FAQ about where was this
1:35:20
discussed before which I'll put a link in the chat too so it I don't want to
1:35:27
give you the impression that this meeting is the first or only time that
1:35:32
this project was discussed but there's always the feeling from it at an
1:35:38
individual level of the first time someone hears about something that's that they think
1:35:45
why wasn't I told about it before um I'm gonna put a link in the chat
1:35:53
if I can find it okay to go um to the
1:35:59
the previous rounds of conversation this one is deliberately more restrictive because the nature of the presentation
1:36:05
is is more restrictive and uh so the nature of the update is more restrictive it's an interim Finance update but I
1:36:12
would point out that the very first employee of the Wikimedia foundation in 2010 was employed
1:36:21
for the purposes of building a paid API which is now long since then that was
1:36:27
Brian Viber long since uh closed as as an API but that helped the foundation
1:36:33
start so the concept of how do we deal with paid API Services is a very
1:36:38
long-standing one inside this inside this uh Community now we're just doing a
1:36:43
much more structured way uh this call is now an hour and 45
1:36:50
minutes long so I think that's probably enough uh for most people
1:36:56
um unless there was any final questions that are directly to the Enterprise project and finance update not Wikipedia
1:37:04
foundation in general
1:37:11
anything further
1:37:17
um I'll just say quickly that uh as someone who is in that group of this
1:37:24
isn't something that I was aware was even under consideration until about
1:37:31
an hour ago uh this has been very helpful and uh I appreciate taking the
1:37:39
time to answer what while I was here felt like exclusively
1:37:45
my questions um I appreciate your appreciation I would recommend you read that last link
1:37:51
I put in the chat too where has this paper been discussed because it links to the major kind of Milestones of this
1:38:00
project as it went from Theory to practice and notably some external blog
1:38:06
posts particularly the one from the open Future Institute that one I think is most important
1:38:12
because it's not written by us so therefore it's a bit more I mean it comes from an organized uh
1:38:18
friend of the family of the Wikipedia movement in open knowledge in the European um public policy space talking about how
1:38:25
this fits in the ecosystem of of reuse of open knowledge data sets and
1:38:31
how to do that in an appropriate way especially in a commercial environment so that's uh possibly a bit more neutral
1:38:37
point of view than just something that we have written ourselves Amy
1:38:43
um yeah sorry if um this was already said I had to take a quick call there um a few minutes ago but um Andrew and
1:38:49
case Liam didn't um already mention it this is being recorded and will be posted
1:38:55
um the whole the whole call so everything that you missed before you joined us so you can get more context
1:39:00
for all of it and there might have been some more information that was shared that will be Illuminating for you so
1:39:06
um uh you can go back and watch the record the whole recording if you'd like as well
1:39:12
I have tried to take time code notes but I didn't start my stopwatch so I I might got I might have got the uh the time
1:39:18
codes at all and that'll be on our project on our project page Liam I look
1:39:23
at this on our main page on meta which media Enterprise on meta there's also the videos of the previous versions of
1:39:29
this call um and I will link it from the talk page
1:39:34
where we said there is a finance announcement and there will be a call I will link the video into that as well
1:39:41
um yep if anyone actually watches an hour and 45 minutes of this call I will be impressed and hello to you in the future
1:39:49
if you did with that I think I will turn off the recording if that's all right with
1:39:55
everyone