Wikimedia monthly activities meetings/Quarterly reviews/Fundraising/January 2015
Please keep in mind that these minutes are mostly a rough transcript of what was said at the meeting, rather than a source of authoritative information. Consider referring to the presentation slides, blog posts, press releases and other official material. Please note: the numbers presented are preliminary as donations are still settling and the fundraising team is in the process of reconciling with the finance department. A report from the entire 2014-2015 fiscal year will be posted after July 2015.
Present: Katie Horn, Anne Gomez, Megan Hernandez, Lila Tretikov, Caitlin Virtue, Caitlin Cogdill, Pats Pena, Tilman Bayer (taking minutes), Lisa Gruwell, Erik Moeller, Damon Sicore, Ellery Wulczyn, Toby Negrin
Participating remotely: none
we area a cross-functional team:
- Major Gifts and [donating] Foundations
- Online fundraising
- Fundraising ops
- Fundraising tech (Anne as link between FR tech and FR ops)
- Ellery does Analytics support
This is the first time the fundraising team has had to meet an increased fundraising goal when traffic –specifically pageviews – are in decline in key fundraising countries. Pageviews are down 8.6 percent in the United States this year compared to last year. However, the budget we had to raise was 16.8 percent higher than last year. So, we needed to step it up.
We reached annual goal 6 months ahead of schedule and we're in a strong position in case traffic declines further next year.
What we planned & did
December year-end campaign
3rd goal: planned to include one payment method that wasn't completed
Tiger team: ...
Erik: are we working with a recruiting firm on that?
Lisa: no, inhouse . . . the recruiting team has found a lot of great candidates
Megan: The France campaign was successful, learned a lot, got good baseline
December online campaign
made gains in each country
(Megan:) Again bulk came from desktop, but massive gains on mobile
Decided to launch desktop before mobile (conversion is much higher on desktop)
this could be done differently next time
Example desktop banners (screenshots)
The large banner was seen only once by each reader
Then, we showed the smaller ones on the second and subsequent impressions to those who had not yet donated
The setup has become more complex than previuous years when we just had "the" banner. This also means that the analytics are also much more complex.
donation rate (over December)
Lila: so that means 6 out of 10 people donated?
Megan: no, ...
60% of donations come from client's first impresson
Lisa: So you can see that there are people whose first impresson was late in the month
Megan: i.e. many donors don't visit frequently
Ellery: number is the probability that impression will lead to donation
Lila: why is probability going down?
Lisa: Later in the month, the less frequent visitors are donating on their first impression. They are less likely to donate than people who visit the site more frequently.
Lila: why rose up again at the end of the month?
Megan: We changed the messaging.
Damon: It is fascinating that we can distringuish between these
could we do analysis based on that?
Lila: we have their email addresses, right?
Ellery, Lisa: yes
Lisa: We also have a donor survey that - remarkably - 10% of donors fill out. We could analyze exising data based on what point in the month they took the survey.
Lila: would need to cut the tail end because of changed messaging
Erik: presumably also (smaller) messaging changes inbetween
Ellery: yes, but mostly it's the daily/weekly periodicity
Damon: do we have data on only the small banner? yes, from last year
Lila, Erik: frequent visitors/power users donate most
Damon: Do we know efficacy of small banner?
Lisa etc.: We can compare with last year and we have A/B tests comparing first impressions of the two banners this year.
Megan: We would not have been able to raise that much money without the large banner in December
Lila: besides impression-based conversion ratio, need also ratio by population
need to understand in which markets we are efficient, where not, and why not
Megan, Lisa: we do study performance relative to traffic and percentage of global GDP and per capita GDP
Toby: is the goal to raise maximum money, or least annoyance?
Damon, Lila: balancing with several dials...
Erik: would you like to highlight some markets based on your data?
Lisa: All of our country campaigns in Q2 met or exceeded the goal. We still have a lot of room for improvement in Japan. We ran that campaign in Q1.
Erik: what is the predominant payment method on Mobile?
Megan: Paypal is most popular
haven't fully cracked the code yet on Mobile, but made good progress
Large Banner Donation Rate (by day/device)
desktop 2-3x better than mobile
Lila: should be only one or two buttons
Damon: compared to other nonprofits' email campaigns, this ratio seems astronomically good
Megan: We have a high open rate. We sent very few emails compared to other nonprofits
CaitlinV: important - these are all former donors already
Damon: can we do comparison per person(?)
Lila: this is good data, but should also have more information how we are trending
e.g. where were we last year on Dec 12 compared to this year on Dec 12
Erik: track on ongoing basis
Lila: what percentage of revenue is recurring?
Lila: that's to be expected, but should optimize that, it's important for our stability
CaitlinV: Major Gifts and Foundations
launched earlier (cognizant of other fundraising efforts in season)
did fundraising events
two $100k donations
work on CRM, want to use more of reporting capabilities
Lisa: we use CiviCRM, and Silverpop for emails
Lila: Silverpop is quite capable
what does Civi give you that Silverpop doesn't?
CaitlinV: we use Civi for major donors, I don't use Silverpop at all
Lila: there are CRMs focused on individuals, and focused on emails
do comparison, intersection can be large
in general: use this time to say where we are going, what do you want do do more, what do you want to do less
Lila: are events worth investment?
CaitlinV: yes I think so, although increasing their number might not be worth it (except perhaps key market like Wikimania country[?])
Toby: I see more opportunities of this kind
Pats: Planning for Q3 and Q4
focused on new countries: China, Brazil (and several other Latin American countries)
talking to vendors there. After China, we'll have time to focus on South America
currently, with international cards can only cover small piece of the pie in Brazil
(Damon, Katie, Lila, Erik, Anne: discussion about working with outside contractors, onboarding, refactoring)
Toby: As a general remark: In Analytics, we have tried to stick with longtime contractors only, because onboarding is so hard
Lila: as we grow, need to get more efficient in that
(Katie:) got dashboard done just in time
Damon: FR tech team can easily get siloed. talking about cycling people in and out of this team
Katie: fixes and upgrades - tried to get all of it done before campaign, but some was done later
During campaign, try to just watch and fix things - system behaves differently under heavy load
in case of problems (e.g with particular payment method) have to stop things as far upstream as possible
Lila: do we have tests?
Katie: some, but they don't catch as much as we'd like
Lila: how often do the alarms ring?
Katie: We did not have any significant problems in December.
Lila: We should improve testing/monitoring.
Toby: cost for migration, hiring?
Katie: We are working to dedup Civi [data].
Lila: usually don't want to do that kind of thing inhouse, but engage contractor instead
first though get the data
goal to update to vendor's new format: in progress
potential (recoverable) data loss within CRM
Lila: that's probably more important for Major Gifts?
CV: yes, Major Gifts already deduped (for >$500). We do that manually
Damon: are you using burndown charts?
Erik: yes, they are the team pioneering the new extension for that on Phabricator
What we learned
Lisa: We are deep in research mode and will release a full report by March 1 on the English fundraiser
We are looking at three areas of analysis:
- What people do (largely Ellery's work and quanitiative)
- What people think: mostly our surveys of both readers and donors, and focus groups (prior and post)
massive amount of data from donor survey (250k participants)
some open ended questions with a lot of responses, e.g. 80,000 people answered "How could we make Wikipedia better for you?" and we are currently sorting and coding these responses
- What people say
donor services team replied to ~25k emails and those are all coded
a social media sentiment analysis is in progress
mailing lists and Meta are being coded
What we learned: technical
when we write documentation, we often write it for ourselves, not sufficient for external contractors - we may need help there
need work to get replica pipeline going
Damon: I understand
Anne: FR goals Q3
Lila: are opinion research results going to come in this q?
Lisa: yes, they are fast
Anne: lot of overlap with tech goals
PCI compliance assessment every year
Lila: more specific about refactoring goals?
Katie: have tasks collected in a project on Phabricator
Toby: do we have mean time for integrating payment vendor?
Katie: no, it varies wildly
Anne: perhaps 2 months at minimum?
Katie: at least
we have data about past ones in Mingle, could look at that, but need to consider that there was lots of pausing due to response time from the vendor
- design support
Lila: you guys need a dedicated designer
Megan: we work with external designers currently
Erik: could you specify this quantitatively? (like "20% of one staff")
Erik: if only occasionally, will ask Jared to provide. otherwise, hire
Lisa: The new reqs for FR tech seem blocked in technicalities. We could use some help moving those along.
Katie: need scrummaster, we are on the edge there
Megan: need to know much more about readers, eg. how often visiting, distribution across devices
Erik: research matrix is very interesting, could be valuable for internal and external conversations about banners, annoyance levels etc.
the March report will be very interesting too, perhaps integrate in a metrics meeting presentation
Lisa: Yes, it might be a good example of being data driven.
Toby: Analytics has signed up fully to support FR
Lila: Ellery did great work