Fundraising 2010/Banner testing

From Meta, a Wikimedia project coordination wiki
In other languages
  Updates       Testing       Translation       Messaging    

Since August, the Fundraising Committee has been running banner & landing page tests. These have evolved from weekly Thursday afternoon tests, which helped us get all of our systems in order and determined which messages would best motivate our donors before the launch of the fundraiser - to almost daily tests introducing new banners and landing pages as we continue to tweak our current campaign.

Check out the Fundraising Updates page where we discuss what we've learned so far and upcoming tests and challenges.

[edit]

December 28[edit]

* We've changed donations over $50 into $50 amounts to account for outliers which may skew our testing data[edit]

Read the update about these tests

$10 vs. $5 test[edit]

If everyone reading this donated $5 vs. 4 days left to make tax-deductible donation test[edit]

Only vs. No Only test[edit]

December 20[edit]

Ask String Test
Read the Update about this test


December 16[edit]

Bold, Bolder, Boldest border test
22:55 - 23:55 UTC

December 15[edit]

0055 - 0155 UTC

Bold Temple Test
22:55 - 23:55 UTC

December 13[edit]

Read the Update about this test
Jimmy Banner Text Test 1855 - 1955 UTC

December 7[edit]

Read the Update about this test
1920 - 2255 UTC

Read the Update about this test
Each banner here is linked to the landing page for this test.
1920 - 2255 UTC

December 6[edit]

Read the Update about this test
2030 - 2255 UTC

0200 - 1755 UTC

December 3[edit]

See the update about this test
Banner Test

Landing Page Test

December 2[edit]

See the update from this test. The Control banner ran at 5%, and editor banners at 47.5%
Editor Banner Test

Landing Page Test

Landing Page Test

November 30[edit]

Kartika Landing Page Test

November 13[edit]

Language Banner Test - Testing for 2 hours on language projects that requested it. First 2 letters of the banner name are the language code. Language banners displayed at 20% Jimmy at 80%. On some wikis small number of relative impressions and visits make rate inadequate.


October 28[edit]

with all data from today's test

October 26[edit]

English[edit]

Spreadsheet with all English data from todays test

Portuguese Wikipedia[edit]

Spreadsheet with all Portuguese data from todays test

French Wikipedia[edit]

Click here for the full spreadsheet with all data from the French test

Dutch Wikipedia[edit]

Click here for the full spreadsheet with all data from the Dutch tests

Swedish Wikipedia[edit]

Click here for the full spreadsheet with all data from the Swedish test

Italian Wikipedia[edit]

Click here for the full spreadsheet with all data from the Italian tests

Italian Wikiquote[edit]

Click here for the full spreadsheet with all data from the French test

October 21[edit]

Spreadsheet with all data for todays test

October 19[edit]

Spreadsheet with all data for todays test

October 14[edit]

Spreadsheet with all data for todays test (interesting!)

October 12[edit]

English Wikipedia[edit]

Spread sheet with full test statistics

French Wikipedia[edit]

spreadsheet with full French statistics

the French tests were run on the 12th of october. The Google spreadsheet does not seem to contain any data at that date. Anthere

German Wikipedia[edit]

Spreadsheet with full German statistics

Test nine: October 11th, 2010[edit]

Text only banners
Full background image banners
Jimmy head background only

Test seven: September 30th, 2010[edit]

Detailed test discussion       stat report      

Test seven Statistics       Spreadsheet with full data (click '09-23 working' at bottom)
Click through rate represents how many people who saw a given banner (an impression) clicked on it. It is calculated by dividing total banner clicks over total banner impressions.

Test six: September 23rd, 2010[edit]

Detailed test discussion       stat report      

Click through rate represents how many people who saw a given banner (an impression) clicked on it. It is calculated by dividing total banner clicks over total banner impressions.
N/A is used for per landing page data because impression data is only useful for the banner as a whole (the TOTAL lines)

Test five: September 16th, 2010[edit]

Detailed test discussion       stat report      

Click through rate represents how many people who saw a given banner (an impression) clicked on it. It is calculated by dividing total banner clicks over total banner impressions.



Test four: September 9th 2010[edit]

Detailed test history       Full stat report       Discuss this test

Banner text % of total clicks
Be curious 28%
Stay curious 24%
Be curious. Be bold. 23%
Who funds {{SITENAME}}? You do. 9%
Think of a world without {{SITENAME}} Keep free knowledge alive. 8%
Who funds {{SITENAME}}? You do. Donate today. 5%
Wikipedia is the world's greatest repository of knowledge. [Donation needed] 4%



Test three: September 2nd 2010[edit]

Detailed test history       Full stat report       Discuss this test

Banner text % of total clicks
Stay curious 60%
Free, Innovative, Non-profit. This is Wikipedia. 14%
The 7th most visited website in the world might be the most important. Donate today. 11%
"If you make one gift this year, make it the gift of knowledge." -- Nelson Mandela 8%
Wikipedia: Supported by readers like you. 7%



Test two: August 26th 2010[edit]

Detailed test history       Full stat report       Discuss this test

Banner text % of total clicks
Help us change the world / Support Free Knowledge 57%
Help us change the world / Donate today to support free knowledge 43%



Test one: August 19th 2010[edit]

Detailed test history       Full stat report       Discuss this test

Banner text % of total clicks
One entry became 33 million. One language became 270.
Help protect what we’ve created.
69%
Wikipedia: Edited by volunteers, supported by readers like you. 31%


Wikimedia fundraisers rely on short banner messages running at the top of project pages. We're looking for messages that catch the attention of readers, and motivate them to click through and donate. The better our banners, the more efficient our fundraising.

Contribute your own banner idea on the Messages page!

Detailed Test Results[edit]

If you want to geek out on the detailed test results click here, check out the in-depth results tables, and join the conversation.