விக்கிப்பீடியா:புதுப்பயனர் போட்டி/Facebook Marketing/Report
Summary
தொகுTamil Wikipedia community is running an article writing contest for new users during Jan - March, 2019.
Through a Facebook marketing campaign supported by the New Readers Team from the Wikimedia Foundation, the Tamil Wikipedia commuinty was able to
- deliver its message to 807,039 people
- engage 299,689 of them
- get 143,644 clicks
leading to
- 86 new users signing up for the contest
- 44 users expressing interest to participate.
The total cost for the campaign was 884.68 USD ( ~62,357 INR)
The campaign ran from December 22, 2018 to January 8, 2019.
The campaign had three phases:
- Phase 1 - To pick the right ad format
- Phase 2 - To pick the right audience
- Phase 3 - To push the best performing message to maximum number of people.
Pick the right ad format
தொகுTamil Wikipedia community had only a novice level understanding of the marketing capabilities of Facebook platform. Therefore, we tried to understand how the ads work and which message performs best. Initially, our focus was on designing different banners and writing catching slogans. We posted them as posts and used the "Boost post" ad option in Facebook. All these were promoted for a day with 10 USD budget each.
In this phase, we learnt:
- Facebook has banner design guidelines which we need to follow. Poorly designed banners with a lot of text won't engage a lot of people and Facebook doesn't run those campaigns effectively. After designing many banner variations, we discovered none of the banners were performing well. Either banners don't engage our audidence or we didn't have the skills or resources to design a banner engaging them and also following Facebook guidelines.
This was a major learning.
And, we discovered that one of the slogans was performing exceptionally well. The slogan format goes like:
"Do you like Animals/ Birds / Movies (insert any field here)? Visit Wikipedia to write articles on them"
We also discovered that Facebook gives many other ad formats besides boosting the posts.
- Get more signups
- Get more website visitors
were two such formats we discovered to drive more clicks to Wikipedia.
We moved to the second phase testing with these learnings:
- Stop boosting Facebook posts
- Niche audience respond better
Pick the right audience
தொகுIn this phase, we ran many different ads for 20 USD each per day per ad.
We tested many different niche audience like Birds, Animals, Science, History, Geography, Movies and Sports.
We created slideshows using Featured Images and Pictures of the Year from the Commons.
These images always performed well irrespective of the supposed popularity of the niche:
- Pictures of the year
- Cute animals and birds pictures
- High resolution pictures
- Color pictures
Black and White pictures we used for testing History niche didn't do well.
Audience demographics
தொகுHaving found the niches that engage better, we also looked at the profiles of the audience responding.
- People between 18 - 44
- Men
- People from all over Tamilnadu, Northern/Eastern/Western Provinces of Sri Lanka
were engaging more with our ads.
So, we continued to test our ads only in these regions of Sri Lanka and not in all the provinces of Sri lanka.
These provinces are predominantly Tamil speaking. Interestingly, Western Province is only the third most populated province for Tamil speaking population. But, it always beat other provinces in engagement. Some times, it also beat a bigger state like Tamil Nadu!
We tested few ads with the age group of 18 - 44 and had a better Click Through Rate but dropped that idea later as a typical contributor for Tamil Wikipedia has come from all age groups in the past.
We also targeted users based on their interests and graudation in certain subjects. For example, for Birds and Animals, we targeted users interested in Wildlife Photography.
Ad quality
தொகுAll the testing above ensured that our ads were super relevant for the audience.
Facebook rated most of our well performing ads 10/10.
In face, some of the ads had more organic comments, likes, and shares.
Ad delivery platform
தொகுWe discovered 100% of the audience came from mobile devices which was a bummer as we were targeting new editors