On 19 August 2015 at 14:26, Sam Klein sjklein@hcs.harvard.edu wrote:
There's a more general problem here we should fix:
We already know that effectiveness of any single banner drops off dramatically after the first few views. So there's rarely a reason to run a continuous banner -- certainly not if there are other banners to run.
I think we should be cautious about using our fundraising experience to predict the efficiency of 'delayed call-to-action' banners like WLM - to my mind they seem to function in quite different roles.
The fundraising banner is calling for an immediate action. You see it, and you either donate or you don't. If you decide not to donate, you probably won't decide to donate on seeing it tomorrow, either; while if you have donated, you're probably not going to donate again. So the banner being repeated doesn't gain us much, and it has progressively less value on the third, fourth, fifth appearances. There are relatively few people who see a fundraising banner and decide "I'll sleep on it", then come back tomorrow and donate. And if they *do*, well - there's a donate link on every page, once they're looking for it.
However, WLM is calling for a delayed action - "go off, do something, and come back again to tell us about it".
The most desired outcome is probably that a previously uninvolved person will see it, click through, think "that sounds fun", and go off to take some photos - after all, it's running all month, they can do it at the weekend. A few days later they come back, and want to upload their photos... but if the banner's not there on Wikipedia, they won't really know where to go. They might not remember the name ("Wiki something?"), making it hard to search for the contest, and they probably didn't bookmark the WLM pages. There isn't anything else on the page that would help to take them there, and if they're not involved in the projects already they probably won't know where the information's likely to be. If we can't make sure they can find WLM easily when they return, then we've wasted the original call to action, we've wasted the potential contributions, *and*, most importantly, we've wasted their time and goodwill.
I think this difference in intended response styles makes it hard to generalise from the "diminishing returns" experienced on fundraising. Yes, a repeated banner will get progressively diminishing clickthroughs. But with WLM, those second clickthroughs in some ways provide the "value" to the first clickthrough - they need to return to make the campaign a success, which isn't really a concern for fundraising. We need to make sure that that channel is open and visible in some way when they come back.
Andrew.