Hi all,
I just wanted to collect some input on the WLM banners.
As background: the centralnotice banners are the most important way that we make newcomers aware of our competition, but it is also the main way many other efforts get attention. This is a scarce resource, that we should use responsibly.
When we started Wiki Loves Monuments, the sitenotice was primarily all-or-nothing, and the community was very much at ease with such an important competition getting this attention. As time progressed, the option to institute a "diet" was added. Unfortunately, the documentation of this is pretty poor, but it looks like this basically constitutes of a maximum number of times that a given device sees the banner within a given span of time. Typically, this number is set at 5 (you only see a banner 5 times), but there is an option to 'reset' that counter to show the banner another 5 times, for example the next week. Some options I would see as feasible (but more experienced sitenotice designers may know more creative solutions) would include: - Show this banner X times in total per device - Show this banner X times per week per device - Show this banner once per day - Show this banner once per Y page views etc
The community has also flagged over the past years that showing the banner without limitation becomes harder to justify as the number of campaigns increases.
Experience teaches us that the effect of a banner diminishes over the number of times it is being shown to people. It may be a small cost to us to determine a diet. For simplicity's sake, we should probably try to come up with one-size-fits-all with some possible local exceptions if special events happen (e.g. a national monument open day).
Now the question is: what is a reasonable and optimal diet to request? This is a tricky balance to strike: what is still enough to almost achieve optimal impact, but minimize the 'cost' in our readers' attention? If people have not clicked on a banner the first five times, will they click on it the sixth?
I invite all national organizers to share their insights here or on the public discussion page. If we can come to a consensus of what we think is a fair diet, then we're also more credible with our request for these essential resources. I have my own thoughts, and I'm sure others on the international team have them too. But first, I'd love to hear some more thoughts. What diet would be fair and practical, if any?
Lodewijk (former international coordination team)
Dear Lodewijk,
In order to answer your question, it would be useful to have information on which diet was eventually implemented during WLM-2019 and 2020. Do you have any such data?
An information on the banner diet during WLE-2021 would be even more useful because all countries with a large number of uploads showed a significant reduction this year compared to 2020. While COVID-19 could play some role here, I suspect that it was not the only culprit.
Sincerely, Alexander
On 8/12/2021 9:53 PM, effe iets anders wrote:
Hi all,
I just wanted to collect some input on the WLM banners.
As background: the centralnotice banners are the most important way that we make newcomers aware of our competition, but it is also the main way many other efforts get attention. This is a scarce resource, that we should use responsibly.
When we started Wiki Loves Monuments, the sitenotice was primarily all-or-nothing, and the community was very much at ease with such an important competition getting this attention. As time progressed, the option to institute a "diet" was added. Unfortunately, the documentation of this is pretty poor, but it looks like this basically constitutes of a maximum number of times that a given device sees the banner within a given span of time. Typically, this number is set at 5 (you only see a banner 5 times), but there is an option to 'reset' that counter to show the banner another 5 times, for example the next week. Some options I would see as feasible (but more experienced sitenotice designers may know more creative solutions) would include:
- Show this banner X times in total per device
- Show this banner X times per week per device
- Show this banner once per day
- Show this banner once per Y page views
etc
The community has also flagged over the past years that showing the banner without limitation becomes harder to justify as the number of campaigns increases.
Experience teaches us that the effect of a banner diminishes over the number of times it is being shown to people. It may be a small cost to us to determine a diet. For simplicity's sake, we should probably try to come up with one-size-fits-all with some possible local exceptions if special events happen (e.g. a national monument open day).
Now the question is: what is a reasonable and optimal diet to request? This is a tricky balance to strike: what is still enough to almost achieve optimal impact, but minimize the 'cost' in our readers' attention? If people have not clicked on a banner the first five times, will they click on it the sixth?
I invite all national organizers to share their insights here or on the public discussion page. If we can come to a consensus of what we think is a fair diet, then we're also more credible with our request for these essential resources. I have my own thoughts, and I'm sure others on the international team have them too. But first, I'd love to hear some more thoughts. What diet would be fair and practical, if any?
Lodewijk (former international coordination team)
Wiki Loves Monuments mailing list To unsubscribe send an email to wikilovesmonuments-leave@lists.wikimedia.org http://www.wikilovesmonuments.org
I actually don't know for sure, but these are good questions. I'm sure that someone (undefined) can look it up who's more familiar with the centralnotice settings.
Lodewijk
On Thu, Aug 12, 2021 at 1:23 PM Alexander Tsirlin altsirlin@gmail.com wrote:
Dear Lodewijk,
In order to answer your question, it would be useful to have information on which diet was eventually implemented during WLM-2019 and 2020. Do you have any such data?
An information on the banner diet during WLE-2021 would be even more useful because all countries with a large number of uploads showed a significant reduction this year compared to 2020. While COVID-19 could play some role here, I suspect that it was not the only culprit.
Sincerely, Alexander
On 8/12/2021 9:53 PM, effe iets anders wrote:
Hi all,
I just wanted to collect some input on the WLM banners.
As background: the centralnotice banners are the most important way that we make newcomers aware of our competition, but it is also the main way many other efforts get attention. This is a scarce resource, that we should use responsibly.
When we started Wiki Loves Monuments, the sitenotice was primarily all-or-nothing, and the community was very much at ease with such an important competition getting this attention. As time progressed, the option to institute a "diet" was added. Unfortunately, the documentation of this is pretty poor, but it looks like this basically constitutes of a maximum number of times that a given device sees the banner within a given span of time. Typically, this number is set at 5 (you only see a banner 5 times), but there is an option to 'reset' that counter to show the banner another 5 times, for example the next week. Some options I would see as feasible (but more experienced sitenotice designers may know more creative solutions) would include:
- Show this banner X times in total per device
- Show this banner X times per week per device
- Show this banner once per day
- Show this banner once per Y page views
etc
The community has also flagged over the past years that showing the banner without limitation becomes harder to justify as the number of campaigns increases.
Experience teaches us that the effect of a banner diminishes over the number of times it is being shown to people. It may be a small cost to us to determine a diet. For simplicity's sake, we should probably try to come up with one-size-fits-all with some possible local exceptions if special events happen (e.g. a national monument open day).
Now the question is: what is a reasonable and optimal diet to request? This is a tricky balance to strike: what is still enough to almost achieve optimal impact, but minimize the 'cost' in our readers' attention? If people have not clicked on a banner the first five times, will they click on it the sixth?
I invite all national organizers to share their insights here or on the public discussion page. If we can come to a consensus of what we think is a fair diet, then we're also more credible with our request for these essential resources. I have my own thoughts, and I'm sure others on the international team have them too. But first, I'd love to hear some more thoughts. What diet would be fair and practical, if any?
Lodewijk (former international coordination team)
Wiki Loves Monuments mailing list To unsubscribe send an email to
wikilovesmonuments-leave@lists.wikimedia.org
-- This email has been checked for viruses by AVG. https://www.avg.com _______________________________________________ Wiki Loves Monuments mailing list To unsubscribe send an email to wikilovesmonuments-leave@lists.wikimedia.org http://www.wikilovesmonuments.org
Great point, Lodewijk
We should ideally be able to measure the relationship between sitenotice and contributors. And, if we currently don't have enough data for that, design WLM in such a way we will be able to measure that so it can be used for future contests.
There is probably a chain of # banner impressions → # visits to contest page → # users actually paying attention → # users participating
While the absolute number of visits to the contest page coming from sitenotice should be relatively simple to count, the number of people that viewed will depend on the campaign configuration and the presence of other campaigns at the same time (e.g. fundraising), plus on the actual traffic received from that country (probably complicated, but could be calculated) *plus* the actual configuration of the user hiding the banner or the diet restricting the number of times it is shown per device.
Then, we have an additional handicap that our reaction won't be immediate.Users are unlikely to have photos of designated Monuments without an image sitting on their computer ready to be uploaded. If a user is going to actually find a suitable nearby monument in the list, plan a trip with its camera (next weekend, perhaps?), actually go out and shoot it, come back, locally review their photos, select good ones and upload them... It will take several days, so the effect of the impressions won't be clear. Furthermore, repeated impressions may be needed to remind them that they intended to actually tak some photos. Whereas most people will simply think "Oh, nice project" and let other people do that work, without wishing to be bothered any more about WLM.
I don't think we actually have the resources to properly measure this, with all its variables. Perhaps we could recruit some help from the Analytics / Product Analytics teams?
Best regards
Hi,
From my perspective as Ukrainian WLM organiser, we would prefer having no diet or, in the worst case, not very restrictive diet.
I understand the arguments of banner blindness. In the Ukrainian community we make sure that there is no other major campaign during Wiki Loves Monuments, due to strategic importance of this contest for our community. Indeed sometimes we might have two or three banners running in parallel in Ukraine, but those are lower priority campaigns with diets.
I think we ran WLM (or it was WLE) with a moderate diet once, and this resulted in frequent cases of potential participants getting lost (e.g. asking us on Facebook what the dates are, or not being able to find the upload link). Note that participation in WLM is not always immediate: people often notice the banner, get the idea (cool, I need to picture a monument!), picture the monument the following weekend and want to upload them — but may fail to find the banner.
Another weakness is alphabet: Ukrainian uses Cyrillic alphabet (Вікі любить пам'ятки / ВЛП), while the international campaign uses Latin alphabet (Wiki Loves Monuments / WLM). Unless you remember exactly the name (which rarely happens when you see the banner exactly once), you will probably take a while to find what you need without a banner.
The worst thing is that any diet disproportionately affects newbies, as experienced Commons contributors know what WLM is and how to find it. And this is really bad for WLM Ukraine, as we have a quite big community of experienced photographers who will participate anyway, but we really want to attract new contributors who never uploaded a picture to Commons (or even never edited any Wikimedia project) before. WLM is quite good at it, please do not make it weaker. If you set a diet, please make it minimal (e.g. show the banner on every second page). It would absolutely not meet its purpose if an uploader has to wait for a week (or several days) to see the banner again.
And of course an A/B test before any change is the best solution.
Mykola (NickK) WLM Ukraine
13 серпня 2021, 01:57:57, від "Platonides" platonides@gmail.com:
Great point, Lodewijk
We should ideally be able to measure the relationship between sitenotice and contributors. And, if we currently don't have enough data for that, design WLM in such a way we will be able to measure that so it can be used for future contests.
There is probably a chain of # banner impressions → # visits to contest page → # users actually paying attention → # users participating
While the absolute number of visits to the contest page coming from sitenotice should be relatively simple to count, the number of people that viewed will depend on the campaign configuration and the presence of other campaigns at the same time (e.g. fundraising), plus on the actual traffic received from that country (probably complicated, but could be calculated) plus the actual configuration of the user hiding the banner or the diet restricting the number of times it is shown per device.
Then, we have an additional handicap that our reaction won't be immediate.Users are unlikely to have photos of designated Monuments without an image sitting on their computer ready to be uploaded. If a user is going to actually find a suitable nearby monument in the list, plan a trip with its camera (next weekend, perhaps?), actually go out and shoot it, come back, locally review their photos, select good ones and upload them... It will take several days, so the effect of the impressions won't be clear. Furthermore, repeated impressions may be needed to remind them that they intended to actually tak some photos. Whereas most people will simply think "Oh, nice project" and let other people do that work, without wishing to be bothered any more about WLM.
I don't think we actually have the resources to properly measure this, with all its variables. Perhaps we could recruit some help from the Analytics / Product Analytics teams?
Best regards _______________________________________________ Wiki Loves Monuments mailing list To unsubscribe send an email to wikilovesmonuments-leave@lists.wikimedia.org http://www.wikilovesmonuments.org
In WLM 2019 and 2020 there was no diet set.
Romaine
Op do 12 aug. 2021 om 22:23 schreef Alexander Tsirlin altsirlin@gmail.com:
Dear Lodewijk,
In order to answer your question, it would be useful to have information on which diet was eventually implemented during WLM-2019 and 2020. Do you have any such data?
An information on the banner diet during WLE-2021 would be even more useful because all countries with a large number of uploads showed a significant reduction this year compared to 2020. While COVID-19 could play some role here, I suspect that it was not the only culprit.
Sincerely, Alexander
On 8/12/2021 9:53 PM, effe iets anders wrote:
Hi all,
I just wanted to collect some input on the WLM banners.
As background: the centralnotice banners are the most important way that we make newcomers aware of our competition, but it is also the main way many other efforts get attention. This is a scarce resource, that we should use responsibly.
When we started Wiki Loves Monuments, the sitenotice was primarily all-or-nothing, and the community was very much at ease with such an important competition getting this attention. As time progressed, the option to institute a "diet" was added. Unfortunately, the documentation of this is pretty poor, but it looks like this basically constitutes of a maximum number of times that a given device sees the banner within a given span of time. Typically, this number is set at 5 (you only see a banner 5 times), but there is an option to 'reset' that counter to show the banner another 5 times, for example the next week. Some options I would see as feasible (but more experienced sitenotice designers may know more creative solutions) would include:
- Show this banner X times in total per device
- Show this banner X times per week per device
- Show this banner once per day
- Show this banner once per Y page views
etc
The community has also flagged over the past years that showing the banner without limitation becomes harder to justify as the number of campaigns increases.
Experience teaches us that the effect of a banner diminishes over the number of times it is being shown to people. It may be a small cost to us to determine a diet. For simplicity's sake, we should probably try to come up with one-size-fits-all with some possible local exceptions if special events happen (e.g. a national monument open day).
Now the question is: what is a reasonable and optimal diet to request? This is a tricky balance to strike: what is still enough to almost achieve optimal impact, but minimize the 'cost' in our readers' attention? If people have not clicked on a banner the first five times, will they click on it the sixth?
I invite all national organizers to share their insights here or on the public discussion page. If we can come to a consensus of what we think is a fair diet, then we're also more credible with our request for these essential resources. I have my own thoughts, and I'm sure others on the international team have them too. But first, I'd love to hear some more thoughts. What diet would be fair and practical, if any?
Lodewijk (former international coordination team)
Wiki Loves Monuments mailing list To unsubscribe send an email to
wikilovesmonuments-leave@lists.wikimedia.org
-- This email has been checked for viruses by AVG. https://www.avg.com _______________________________________________ Wiki Loves Monuments mailing list To unsubscribe send an email to wikilovesmonuments-leave@lists.wikimedia.org http://www.wikilovesmonuments.org
As I just wrote on https://meta.wikimedia.org/wiki/CentralNotice/Request/Wiki_Loves_Monuments_2...
As I tried to say earlier on the Meta page, I am trying to make sure a banner works, but decision making around banners should be done with the international in dialogue with the local teams and the broader community. The discussion on the mailing list and here is very very limited in comparison to the number of people involved. This is concerning.Another concern is that the proposal above, to implement a diet with a "cooldown" is I think enlarging an existing problem (banner blindness) and creating a new problem (new participants who can't find back the instructions page, aka less uploads). Let me try to explain. Banner blindness is influenced by for major factors:
1. The first factor is how many different banners a person sees in a certain period. If someone sees 6 banners in a certain period, banner blindness is worse than only seeing one banner in that period. 2. The second factor is how many times a certain banner is visible above the pages. If a banner is only shown five times, it is less blinding than when seen 50 times. If a banner disappears, a lot of users will think this banner is gone (but when a cyclus is in place the annoyance can be larger, see 4). 3. The third factor is on how many different wikis someone is active on which the same banner is shown. If a diet is put in place (let's say 5x shown) and a user has visited above that number of pages, a banner disappears (or after clicking it away), that same banner will re-appear when that user visits another wiki. A diet does not work cross wiki. 4. The fourth factor is how many cycles a banner is shown to a single user. (A cycle means that when someone got a banner a few times, after some time (days/weeks) a banner can be shown again for a few times.) While this is actually the same banner, a lot of users would experience this as a second/third/etc banner, or in other words, the banner blindness is higher.
(There can be also other factors in place, like banners that are shown but are relevant for only an extremely low number of viewers, as well as banner designs (like moving flags) that can create much annoyance, but above I just focussed on the different ways of how the same banner can be shown.) The third factor (in my opinion) should have been fixed already five years ago and is a long standing problem. However I do see some progress, you can now disable certain type of campaigns in your preferences if you do not want to see them: Special:Preferences#mw-prefsection-centralnotice-banners https://meta.wikimedia.org/wiki/Special:Preferences#mw-prefsection-centralnotice-banners . *Concerning the proposal to put a diet in place with repeating cycles:* At first sight this proposal seems to be helpful to reduce banner blindness and annoyance for banners. However, users would not get the banner once, but four times (factor 4), which would cause 4x the annoyance, and if they work cross wiki this would double/triple/etc the amount of banners they experience (factor 3). But more importantly: this diet would cause a major issue: new editors can't find back the instructions page of the photo contest when they re-visit Wikipedia (or the page is otherwise much harder to find back). In other words: a lot less photos uploaded by a less diverse amount of uploaders. I am a huge fan of diets, but above all a banner (+diet) must by *suitable* for the goal of the banner. In other words, the WLM banner has 2 goals: 1. to inform people about the contest ongoing, 2. after participants have made photos they return to Wikipedia and then the banner has as goal to enable participants to find back the instructions page. The proposed diet is obstructing the second goal and is damaging for the photo contest.
So if you as local organisers do not wish a bad effect because a diet is in place, replying on the Meta page and/or here on the mailing list would be needed.
Romaine
Op di 31 aug. 2021 om 03:56 schreef Romaine Wiki romaine.wiki@gmail.com:
In WLM 2019 and 2020 there was no diet set.
Romaine
Op do 12 aug. 2021 om 22:23 schreef Alexander Tsirlin <altsirlin@gmail.com
:
Dear Lodewijk,
In order to answer your question, it would be useful to have information on which diet was eventually implemented during WLM-2019 and 2020. Do you have any such data?
An information on the banner diet during WLE-2021 would be even more useful because all countries with a large number of uploads showed a significant reduction this year compared to 2020. While COVID-19 could play some role here, I suspect that it was not the only culprit.
Sincerely, Alexander
On 8/12/2021 9:53 PM, effe iets anders wrote:
Hi all,
I just wanted to collect some input on the WLM banners.
As background: the centralnotice banners are the most important way
that
we make newcomers aware of our competition, but it is also the main way many other efforts get attention. This is a scarce resource, that we should use responsibly.
When we started Wiki Loves Monuments, the sitenotice was primarily all-or-nothing, and the community was very much at ease with such an important competition getting this attention. As time progressed, the option to institute a "diet" was added. Unfortunately, the
documentation
of this is pretty poor, but it looks like this basically constitutes of a maximum number of times that a given device sees the banner within a given span of time. Typically, this number is set at 5 (you only see a banner 5 times), but there is an option to 'reset' that counter to show the banner another 5 times, for example the next week. Some options I would see as feasible (but more experienced sitenotice designers may know more creative solutions) would include:
- Show this banner X times in total per device
- Show this banner X times per week per device
- Show this banner once per day
- Show this banner once per Y page views
etc
The community has also flagged over the past years that showing the banner without limitation becomes harder to justify as the number of campaigns increases.
Experience teaches us that the effect of a banner diminishes over the number of times it is being shown to people. It may be a small cost to us to determine a diet. For simplicity's sake, we should probably try
to
come up with one-size-fits-all with some possible local exceptions if special events happen (e.g. a national monument open day).
Now the question is: what is a reasonable and optimal diet to request? This is a tricky balance to strike: what is still enough to almost achieve optimal impact, but minimize the 'cost' in our readers' attention? If people have not clicked on a banner the first five times, will they click on it the sixth?
I invite all national organizers to share their insights here or on the public discussion page. If we can come to a consensus of what we think is a fair diet, then we're also more credible with our request for
these
essential resources. I have my own thoughts, and I'm sure others on the international team have them too. But first, I'd love to hear some more thoughts. What diet would be fair and practical, if any?
Lodewijk (former international coordination team)
Wiki Loves Monuments mailing list To unsubscribe send an email to
wikilovesmonuments-leave@lists.wikimedia.org
-- This email has been checked for viruses by AVG. https://www.avg.com _______________________________________________ Wiki Loves Monuments mailing list To unsubscribe send an email to wikilovesmonuments-leave@lists.wikimedia.org http://www.wikilovesmonuments.org
wikilovesmonuments@lists.wikimedia.org