Wikipedia talk:Fundraising/2022 banners

Thank you to WMF, and a little reality check

First, thank you to the WMF folks who listened to the concerns here and create this page for workshopping banner ideas.

I would also like to inject a little caution to our editors. I'm far from an expert on either fund-raising or marketing, but I've done a little of both (and had the good fortune to work closely with people who were indeed experts at it). One of the things I learned is that real-world testing is critical.

A common technique in planning a marketing campaign is what we're doing here on this page; sit around a big table, toss out message ideas, and hash them over. Then pick, say, the 10 that you think are the best. The next step is to run a small-scale trial. Use a small fraction of your marketing budget to get all 10 out in the field and see which ones work and which don't. Every time I've been involved in this, it's been an eye-opener. WTF? Only 1% of readers clicked on X, but 4% clicked on Y? How is that possible? X is obviously the better one! Well sorry, that's not what our audience says.

So please don't take this as a vote on which banners are going to run. -- RoySmith (talk) 15:31, 25 November 2022 (UTC)

Yes, thank you, WMF! I'm still going to scroll past them, but they look so much better IMHO. I'm glad this apparently all got worked out. I'll keep the donations template on standby at the help desk/Teahouse, but maybe we'll get less confused/upset drive-by posters this year. Maybe. 199.208.172.35/97.113.27.216 (talk) 15:50, 25 November 2022 (UTC)

@Roy I'm guessing you might already know this but for those who don't, this is normally what they do. In fact they "did" it for this year's campaign. Because of the increase in chartiable giving during this time of the year I'm guessing that this will be done while the campaign is live rather than honing ahead of time. Best, Barkeep49 (talk) 16:46, 25 November 2022 (UTC)

Yes, I'm sure the WMF folks do that. I was just pointing out to the other people participating that "help us come up with ideas" is not the same as "vote on which ideas to use". -- RoySmith (talk) 16:57, 25 November 2022 (UTC)

Related stats and tests

2 things I would be overjoyed to see:

• A rough quantification of the size + emotional valence + net intrusion factor of banners. Then we could look at the marginal benefits of the largest / most alarming banners, and the best performing uplifting ones (with a negative net intrusion score?)

• Experiments in banners driving account-creation and editing. I don't think we've done this for a long time, despite those being core metrics of community health.

It's nice to see end of year messages that mention contributing $ and edits in the same message, how much better if those tests could.be grounded in what we've found are the most uplifting calls to do each independently. – SJ + 16:20, 26 November 2022 (UTC)

Thanks SJ. We do try and keep a close track of banner size. However quantifying other factors of intrusion has always been much harder. For just one example, we've encountered widely differing views about whether banners with a blue background or a white background with red border are more intrusive.

Regarding editing and account-creation banners, I think you'll be interested in Marshall's post here about recent experiments. Peter Coombe (WMF) (talk) 18:08, 26 November 2022 (UTC)

Thank you! In terms of size, do you track size proportional to screen size on mobile? I like them much better they fit on one page, and am curious that current mobile-large is around 1.5 pages (at modest screen size + font size). I'm sure for some people it is 2x that. If you bucket responses by mobile screen size, do you see a difference in how large banners perform? – SJ + 22:07, 29 November 2022 (UTC)

SJ’s request for a quantification of the effect of "emotional valence" might be hard to perform, however I believe a comparison of banner #2 (which clearly says the WMF is financially stable) with some reference (the best-performing banner of this year?) would be important. It should be easily feasible. I realize a clickthrough ratio with three decimal places might be more than the WMF is ready to share but a ballpark figure ("about 20% less effective") would be nice. Tigraan^{Click here for my talk page ("private" contact)} 18:41, 29 November 2022 (UTC)

+1. And if it simply doesn't work ("about 80% less effective") that would also be good to know. – SJ + 22:07, 29 November 2022 (UTC)

Hi everyone,

We tested banner heights in the past but this week we're focused on finding new messages for the banners so will save this idea for later. You can also find an update on how the banners performed on the first day here. Best, JBrungs (WMF) (talk) 11:50, 1 December 2022 (UTC)

@JBrungs (WMF): Thanks for the first-day update. It says Banner 2 was trending behind banner 1, and banner 3 and 4 brought in between 15-25% fewer donations than banner 1. Can we have the actual figure for banner 2 vs. banner 1? "Trending behind" could be -5% or -95%, but those entail very different conclusions. If necessary, round to the closest 5% and/or add a disclaimer about methodology, confidence intervals etc. Tigraan^{Click here for my talk page ("private" contact)} 12:41, 1 December 2022 (UTC)

NOBANNER

Does WP:NOBANNER still apply? I have "Fundraising" enabled in preferences but logged in do not see a banner. Checked with chromium no extensions. fiveby(zero) 18:13, 30 November 2022 (UTC)

@Fiveby the campaigns currently include the "fundraising" tag, but the banners also have the "logged in users" display set to off (example). The easiest way to see them as a reader would be to open a page in a private browsing session. — xaosflux ^Talk 18:27, 30 November 2022 (UTC)

Thanks, mostly sure they have been visible in past years. A preference checkbox that doesn't allow me to express my preference should probably go away. fiveby(zero) 18:41, 30 November 2022 (UTC)

It is still useful, should a campaign that also targets logged in users be launched. I think that right now the WMF is avoiding targeting individual editors who are mostly already donating their time and talent. — xaosflux ^Talk 19:05, 30 November 2022 (UTC)

There are some small affiliate tax campaigns that do get shown to logged in users so it does serve a purpose. Seddon ^talk 04:00, 30 December 2022 (UTC)

Suggestion on fundraising period

The local classical station I listen to has multiple fundraising campaigns over the calendar year. One thing they practice worth considering here is that their campaign lasts only until the scheduled period ends or they meet their stated goal first. They do this, IIRC, because they see themselves as a non-profit whose mission is not to raise money, but to do things like promote the enjoyment of classical music as well as run a radio station. I believe that it would better fit the mission of the Wikimedia Foundation if they had a similar approach: state how much money they intend to raise, then end the campaign if this goal is met before the scheduled period ends. The Foundation has an endowment, so there is no risk should the time run out before the goal is met. -- llywrch (talk) 21:09, 17 December 2022 (UTC)

The WMF has stated a goal, at least on this page, and they have implied in some of their proposed messaging that if everyone donated $2, they could end their fundraiser today, so I think your request is being met. The only thing they need to change is their fundraising goal, which may need to be lowered to account for the decreased return from messaging that is more honest. – Jonesey95 (talk) 21:36, 17 December 2022 (UTC)

@Jonesey95 and Llywrch: The WMF has for the past few years had the goal of maximising its revenue. [1] In 2018, the WMF's Revenue Streams group reportedly "estimated that coming closer to our vision would probably require an annual budget for the movement in the vicinity of a billion dollars". This is about six times more than the Foundation has taken in recent years. As far as the WMF is concerned, the sky is the limit.

What this has meant in practice is that the WMF hasn't viewed fundraising goals as targets to meet, but as targets to beat. In 2020-2021, for example, the WMF started out with a fundraising goal of $108 million for the Foundation (plus $5 million to be raised directly for the Endowment). Fundraising went well, and in the second quarter the Foundation target for the year was raised to $125 million. But by the end of the third quarter, even this revised target had been exceeded, with the Foundation posting a third-quarter total of $142 million (which is when it started fundraising in India and South America), and by the end of the fourth quarter it had taken $154 million – with the $108 million goal it had started the year with long forgotten. You can see the progression here: Q1, Q2, Q3, Q4. (Note that the Endowment target was also exceeded, by $13.9 million. Also note that these are the sorts of slides the WMF has now stopped publishing ...)

So a wording like "our fundraiser would be over if ..." – absent a clear and public statement of a change in revenue policy such as Llywrch suggests – is disingenuous. Andreas JN466 09:56, 22 December 2022 (UTC)

The endowment is actually a rather limited safety net given that (1) the whole point of an endowment is that the principal should never be dipped into and (2) the income from a $115 million endowment is very modest compared to the WMF's spending. (At 3%, the endowment would yield just about $3.5 million a year, covering a mere 2% of this year's planned expenditure.) Andreas JN466 14:00, 22 December 2022 (UTC)

The issue of endowments is that their revenue rarely, if ever, covers the yearly budget of a given non-profit. It's more of an insurance policy against a bad fundraising year. And I have this impression that the goal of the endowment was that should the projects reach some point where there aren't enough volunteers to maintain them, & the best solution was to simply put them in read-only mode as an online archive, that was what the endowment would fund. If I am right, then $3.5 million would adequately cover yearly server, hosting & networking costs. (You'd need to fund some kind of skeleton staff to keep the physical machines running, apply security patches to the server software, & handle the necessary overhead.) Otherwise, the endowment is nothing more than a pot of money people receive money to oversee -- viz. a sinecure for a lucky few. -- llywrch (talk) 19:32, 22 December 2022 (UTC)