Article feedback/Research/Call to action

Jump to: navigation, search
For a discussion of future development of Calls to action, see this page

July 2011[edit]

A summary of the analysis of Call-to-Action data as of July 2011 can be downloaded here.

The following is an update of CTA data analysis performed in July 2011 and based on a 30-day period observation (2011-06-04 to 2011-07-04) of a random sample of ca. 100K articles of the English Wikipedia, with CTAs tracked at 10%. The click-through/completion rates are consistent with those reported in May 2011 and based on data collected via AFT v.3. This new analysis adds separate click-through rates for Signup and Login events in the Join CTA. Completion and conversion rates are only available for Edit and Survey CTAs.

CTA click-through, completion and conversion rates (v.4)
Edit 17.4% 15.3% 2.7%
Join (signup) 4.7% n/a n/a
Join (login) 3.5% n/a n/a
Survey 37.2% 64.8% 24.1%
Total and hourly volume of events for different CTA (v.4)
CTA table abs 8.pdf CTA table volume 8.pdf

May 2011[edit]

A summary of the analysis of Call-to-Action data as of May 2011 can be downloaded here.

Starting with Version 2 of the Article Feedback Tool (deployed on March 14, 2011 and applied to a sample of approximately 3,000 articles as of March 28, 2011), a series of calls to action (CTA) have been introduced, prompting users to engage in specific kinds of action right after submitting their ratings. The three CTA currently implemented invite users to edit the article, to create an account or to take a survey after rating an article.

We collected data during two short periods (v.2: 2011-04-24 – 2011-04-28; v.3: 2011-04-29 – 2011-05-06) and observed click-through rates and the volume of successfully completed actions. The completion of account creation CTA was not tracked due to technical limitations.

Hourly click-through rates for different CTA (v.3) [blue: accept; red: reject]
AAP cta edit plot ctr 3.pdf AAP cta join plot ctr 3.pdf AAP cta survey plot ctr 3.pdf

The results show that not only users who rated an article have a high likelihood to take a follow-up survey (click-through rate: 40.1–40.6%), but also to successfully complete it (completion rate: 66.3–68.6%). Results are also promising in the case of the edit CTA where 1 out of 6 users who rate an article accept the invitation to edit it (click-through rate: 14.8–15.1%). The actual completion rate for editing remains to be determined (0.8–17.0%) but it is significantly lower than the completion rate for surveys, possibly due to the higher technical barriers to complete the task or the lower relevance of the call after the feedback form. The account creation option generated by far the lowest number of clicks among the three CTA (3.6–4.4%) suggesting that it is highly unlikely that users who submit ratings can be easily persuaded to create an account via a simple CTA mechanism such as the one currently deployed.

CTA click-through and completion rates
CTR v.2 CTR v.3 CMP v.2 CMP v.3
Edit 14.8% 15.1% 0.8% 17.0%
Join 4.4% 3.6% n/a n/a
Survey 40.1% 40.6% 68.6% 66.3%
Total and hourly volume of events for different CTA (v.3)
AAP cta table abs 3.pdf AAP cta table volume 3.pdf