Talk:Fundraising/2019-20 Report

From Meta, a Wikimedia project coordination wiki

Thank you for reading the report. We are happy to share information about fundraising at the Wikimedia Foundation and look forward to your feedback. We will respond to comments and questions to the best of our ability. In some cases, we may not be able to provide all data requested for a number of possible reasons, including the privacy and protection of our users and donors, security considerations, data limitations, and the team's capacity.


«This year, our mobile fundraising banners had an improved performance: an increase of 14 percent in mobile banner revenue and 25 percent in mobile banner donations». Revenues are not a measure of performance, because they don't include a consideration of costs. To say that performance has increased, you'd need at very least to compare that with the number of impressions multiplied by average size of the banner, otherwise we have no idea how much screen estate has been consumed in the process. Ideally we'd also be able to measure other costs.

Same for email: alright, the first email brought more income, but what was in it? How much did you turn on the heat with its contents? How many other threatening emails have been sent? Etc. Nemo 18:02, 6 October 2020 (UTC)[reply]

They stopped including the email count in the 2017-18 report. ("We felt that the raw number wasn't helping in providing any real insight or context into our activities.") Based on the previous rate of continual growth of the number of emails, it's fairly likely that the number has still been steadily increasing. We really should also have a list of the contents of all previous fundraising emails the WMF has ever sent out. We could probably scrape this together ourselves, but the WMF really should be providing it. --Yair rand (talk) 22:50, 6 October 2020 (UTC)[reply]

Minor point[edit]

Hi. Thanks for sharing this document. Is there a reason why Central America and the Caribbean are included in North America on the report map? --Joalpe (talk) 16:09, 13 October 2020 (UTC)[reply]

@Joalpe: Most of the English-speaking world, unlike many other places, generally divides the continents such that those areas are included in North America. (See, eg w:North America.) --Yair rand (talk) 23:56, 13 October 2020 (UTC)[reply]

No response to issues raised by contributors[edit]

While the continual increase in revenue is impressive, this report feels incomplete, as it does not talk about the cost of fundraising, financial and goodwill. By that, I mean:

  • direct costs for the fundraising team
  • number of impressions on desktop and mobile
  • number of emails sent
  • any change in consumption patterns for users shown banners

There have been a number of complaints on the mailing lists about these items and imnit touching them is disappointing. Strainu (talk) 18:59, 14 October 2020 (UTC)[reply]

The total budget for fundraising for FY19-20 was $12.5 million. That includes the cost of payment processing, staff, overhead and all other costs associated with fundraising. The audited actual expense will be published in WMF's annual report later this year.--Lgruwell-WMF (talk) 21:36, 16 October 2020 (UTC)[reply]
In addition, these costs will also be further detailed in the Wikimedia Foundations Form 990. Seddon (WMF) (talk) 01:03, 17 October 2020 (UTC)[reply]
A few years ago we documented in a blogpost a significant change to our fundraising campaigns where we brought in impression limiting. The result of that change was to generally limit banners to 10 impressions per browser and following that are suppressed for 30 days which is the typical length of a major fundraising campaign. We broadly monitor pageviews and impressions for significant divergences but we don't do this sort of deep dive report year on year focused on impression data. Our approach in our biggest campaigns where we serve the bulk of our impressions hasn’t changed though and you can see this in the current campaign settings for our french campaign.
Similarly for emails, our email list grows and with that the total number of emails grows but again it's not really an effective measure for how intense an email campaign is. The more accurate measure for intensity for email is how many fundraising emails we send per person.. For the vast majority of our donors this is limited to 3 fundraising emails in a given year. This has remained in place for at least the last 4-5 years. Seddon (WMF) (talk) 15:49, 19 October 2020 (UTC)[reply]
The total number presumably would be helpful for measuring how effective the campaign is per-email, and for measuring disruption on an absolute level, but yes, ideally we should have both complete totals and per-person totals.
I agree with Strainu that all of these elements should be published. --Yair rand (talk) 21:52, 22 October 2020 (UTC)[reply]

Chapter gifts[edit]

This is such a good report format; appreciated. How are Chapter Gifts defined now? I don't think it is defined anywhere here or on meta. That would make a good footnote for future iterations. (Seddon or anyone who can say) –SJ talk  00:40, 26 July 2021 (UTC)[reply]

South America + Africa numbers[edit]

Hi @Seddon (WMF): and all, Can you provide more detail on banners + email outreach in South America, Africa, and Asia?

The Asia + Africa totals are surprisingly low, given their populations and GP, even given that our reach there per capita is lower than in the rest of the world.

And the South American total is vastly lower, despite reasonable reach. I heard that banners were not run there due to COVID during 2019-20. But were there not low-volume test banners at other times of the year? And the total for 2018-19 was also quite low. Is there some other explanation for it? –SJ talk  23:00, 3 November 2021 (UTC)[reply]

Hi SJ,
Thank you for your questions. We have since published our new Fundraising report for 2020/21. Have a look over there for updated regional totals from last year.
While we run fundraising campaigns each year in most of our fundraising countries, we do adjust campaign schedules for a variety of reasons including local scheduling optimization, accommodating Wikimedia community banner campaigns, regional circumstances, and team capacity. Some countries did not receive a full fundraising campaign in FY20-21 (often campaigns straddle two fiscal years). Other countries did have campaigns in FY20-21, but not in the previous year, which led to revenue differences year to year. This year was also the first year running a campaign in India, which led to an increase in revenue in the region.
We ran banners in Q3/4 (April and May) in several Latin American countries. Due to community campaigns, we did however shorten our campaigns to accommodate these local campaigns (eg. WikiLovesMexico, Spanish Wikipedia’s 20th birthday, etc.). We also decided not to run banners in Colombia due to the political unrest the country was experiencing at the time.  
Just a FYI, Seddon has moved on to a different role in the Foundation but don’t worry, you can contact me for any questions you might have. JBrungs (WMF) (talk) 08:25, 5 November 2021 (UTC)[reply]