Research talk:Wikipedia Banner Challenge

Survey questions[edit]

Hi, this sounds like an interesting idea, can you give me a link to the question set please. WereSpielChequers 23:47, 15 December 2011 (UTC)[reply]

Sure, you can see a beta version here: http://staging.allourideas.org/wikipedia-banner-challenge The website combines two main elements: 1) pairwise voting to sort banners and 2) uploading of new banners that are then sorted by others. Through this process, we both collect and prioritize banner ideas. In the end, we have the overall sense of the community about which banners are most promising. All uploaded banners will be reviewed before going to live site. Msalganik 00:00, 16 December 2011 (UTC)[reply]

Other info[edit]

Are you intending to collect any information other than the question answers, for example value of donation, edit count, tenure etc. If so you will need to disclose that up front, especially if you include something potentially unique or which in combination could identify the subject such as edit count or userrights and account creation date. WereSpielChequers 23:47, 15 December 2011 (UTC)[reply]

No, we are not planning on collecting any of this information, or any other information, that might be potentially identifying. Msalganik 00:02, 16 December 2011 (UTC)[reply]

Thanks, in one sense that's reassuring. But what happens with attribution of banner suggestions? Wikimedia has a culture and license involving attribution, but unless I've missed something there is no way to attribute those banner suggestions to the authors, or anything informing people about the licensing status of their contributions. You really need to tell people if you are switching them from a site that gives attribution for content to one that doesn't.

I may have spotted an exception to the information collection bit. In order to post to the forum you have to go through a Google groups process which in my case tried to out my real name. WereSpielChequers 01:33, 16 December 2011 (UTC)[reply]

In terms of attribution, it is correct that we have no way to attribute the banner text, but the banner images are all either from the Wikimedia Foundation or are in the public domain. Not collecting information for attribution is part of how we avoid collecting any identifying information. In terms of the Google Group, that is not required for participation in the Wikipedia Banner Challenge. We set up the Google group so that people could report bugs or make suggestions in a way that other people can see and build on. However, we also do take bugs and suggestions by email. Msalganik 22:23, 16 December 2011 (UTC)[reply]

Yes I've no issue with the voting between different banners being unattributed. But it does seem odd to me that we are seeking unattributable banner suggestions, and I didn't notice anything that pointed this out to people. WereSpielChequers 22:40, 16 December 2011 (UTC)[reply]

Integrity[edit]

Because this uses photos of genuine Wikimedians but with different messages, potentially each others messages or even words you put in their mouth, I think this compromises the integrity of the message that has just prompted our donors to donate money. If anyone can put any words in Brandon's mouth you are bound to wonder if the message from Brandon that you just responded to was authentic. I don't think I'd have the same concern if the pictures had been omitted or were consistently associated with the people who said them. But for the longterm support of our donors I would suggest omitting the photos or only using them with the words said by that person. WereSpielChequers 01:45, 16 December 2011 (UTC)[reply]

I agree that integrity is very important to the fundraising campaign, and this project would not hinder that. First, there are two main components of the donation process, the banner and the appeal (e.g., http://wikimediafoundation.org/wiki/L11_3009_Brandon/en/US). The Wikipedia Banner Challenge does not in any way involve the appeal. That would still be coming straight from Brandon and the other Wikipedians. Also, the Fundraising team is not required to use any banner that is not consistent with their policies or with the overall spirit of the campaign. In other words, I think the goal of the Wikipedia Banner Challenge is to help the community generate ideas for the Fundraising team. The Fundraising team then has to use judgement about what actions they should take based on those ideas.Msalganik 23:18, 16 December 2011 (UTC)[reply]

Yes but the donors may not realise that the appeal is genuinely from the person whose name is on it, especially if the calls to action are shown to be unrelated. That might not matter if the calls to action were all obviously generic at the point when the donors first see them, but they aren't. So in my opinion there is a risk that our core donors could perceive us as acting without integrity, and while that probably won't affect this years donations it could hurt us in the future. WereSpielChequers 12:09, 19 December 2011 (UTC)[reply]

Please click on the Wikipedia fundraising banner that makes you want to donate more.[edit]

resolved

I'd like to suggest changing this call to action. Currently it is ambiguous between three meanings:

Please click the banner that you think would persuade more people to give money to Wikimedia.
Please click the banner that you think would persuade donors to make larger donations to Wikimedia.
You've just given us a little money and we'd like to persuade you to give a bit more.

I'm assuming that we are happy with something that is ambiguous between the first two but the third is an unintentional side effect of using this on brand new donors as part of the thankyou for their donation. Would Please click the banner that you think would best persuade people like you to give money to Wikimedia. be an acceptable alternative? WereSpielChequers 08:48, 16 December 2011 (UTC)[reply]

This is a tricky issue and one we considered carefully. Based on our user-testing, we opted for "Please click on the Wikipedia fundraising banner that makes you want to donate more" because our testers reported that this was clear and simple. We decided not to ask the participants about people like you or other people because we did not want to ask them to guess about the behavior of others. Ultimately, the choice of the best text is really an empirical question, but one for which we lack the data we need. We could experimentally try several different call-to-action/instructions and then see in which case the scores from Wikipedia Banner Challenge are most predictive of donation behavior during the fundraiser. Unfortunately, we have not done such an experiment yet, so we had to rely on the results of our user-testing. I hope we will be able to do experiments like this in the future. Msalganik 23:28, 16 December 2011 (UTC)[reply]

Thanks for the explanation. I assume your user testing wasn't of people who'd just completed a donation? I wouldn't be quibbling about this wording except for the issue I raised about some donors seeing this as a sneaky way to persuade them personally to give more. I take your point about not wanting to ask people to guess the behaviour of others, may I suggest Please click the best of these two banners. WereSpielChequers 23:44, 16 December 2011 (UTC)[reply]

Interesting point. No our testers had not just donated. I do see how they could interpret the situation differently. There will be a description of the project on the thank you page near the link. That description will have to be clear that we are not trying to convince them to donate again. Also, initially we tried Please click on the better banner. and participants asked us what we mean by best. That is what caused us to move to the more concrete question.Msalganik 23:12, 19 December 2011 (UTC)[reply]

OK as long as you find a way to make it clear that this is about choosing which ads to show to others and not about getting them to increase their donation then I'm satisfied. Otherwise, as you point out, the wording needs to be based on what works after testing, and if my preferred wording doesn't work then there's no point trying it. WereSpielChequers 01:48, 20 December 2011 (UTC)[reply]