Have you considered the possibility that maybe some users just have www.wikipedia.org as their homepage? If that were the case, I would expect most of their (rather frequent) visits to end up in them clicking a bookmark or going to the address bar.

(I don't know how the 43% number was measured, but if it's unique visitors, then obviously my hypothesis is incorrect.)

Hi @Yair_rand - yup, we definitely have considered the possiblity that the wikipedia.org portal page is set as the home page for many browsers. Unfortunately, we have no way to determine how many browsers are setup in this manner.

This ticket is to meant to create (or otherwise use) a type of survey tool to see if we can capture some feedback from the users that "see" the portal page but don't actually use it.

Drafted sample survey to be used on Qualtrics using these (sample) questions:

1. Why did you decide to come to Wikipedia.org?
2. Did you find what you were looking for when you came to Wikipedia.org?
3. What language(s) do you prefer to read or search in?
4. Is there something you'd like to see on this page that would be helpful to you?

There will also be a thank you note, privacy policy links (WMF and Qualtrics) and a section where a visitor can give us their name and email to talk further about the portal page (if they desire).

In T129362#2218037, @debt wrote:

Drafted sample survey to be used on Qualtrics using these (sample) questions:

1. Why did you decide to come to Wikipedia.org?
2. Did you find what you were looking for when you came to Wikipedia.org?
3. What language(s) do you prefer to read or search in?
4. Is there something you'd like to see on this page that would be helpful to you?

There will also be a thank you note, privacy policy links (WMF and Qualtrics) and a section where a visitor can give us their name and email to talk further about the portal page (if they desire).

Did you draft the list of answers too?
In my experience the choice of answers is at least as important as the question itself.

Yes, here's the full (draft) questions and answers via screenshots. Please let me know if there are questions, comments, suggestions...

These questions (and answers) are quick, to the point, and offer up the option to continue the conversation. 👍

Updated survey to be much shorter and pointed in answers, sent to Legal for review and privacy policy generation.

updated banner mock with privacy link and legal language

Legal approval has been achieved and privacy policy has been created. We'll create the banner and link to the WMF privacy policy and the Qualtrics privacy policy as well as the button that will go directly to the survey.

The goal is to get approximately 5,000 to 10,000 responses with this survey. To do this, we will put this banner in front of 1 in every 500 visitors to the site, using the existing EventLogging testing code logic.

We will be able to control the actual access to the survey within Qualtrics, so that we don't exceed the limits of the amount of responses that we have available.

We anticipate that this survey and banner will go live next week - we'll keep it on for at least a day (the banner). Once we're at our desired response rate, we'll turn off the survey from the Qualtrics website and then remove the banner from the portal page shortly thereafter, during a normal SWAT deployment window.

Added information to the Research page:

https://meta.wikimedia.org/wiki/Research:Wikipedia.org_Portal_Survey

and

https://meta.wikimedia.org/wiki/Community_Engagement/Calendar

Updated the banner in the ticket description to be on two lines and added the word 'statement' .

The banner went live on Tuesday, May 10 during morning swat for 1 in every 500 visitors. We updated the script later on that day to hit 1 in every 50 visitors, due to very slow rate of responses.

We had 318 visitors open up the survey as of Tuesday morning (9am MT) on May 17th and decided to change up the algorithm a bit to show to 1 in 30 visitors. This change was done during morning swat and we will monitor it throughout the day.

The survey was turned off on May 17, 2016 and the results compiled here: https://commons.wikimedia.org/wiki/File:Wikipedia_Portal_Survey_-_May_2016.pdf