Firefox/Input/Thank you page: Difference between revisions

From MozillaWiki
< Firefox‎ | Input
Jump to navigation Jump to search
(→‎Phase 1: updated thank you page project phase 1 requirements)
(add another idea for a future phase)
Line 70: Line 70:
Future phases may include:
Future phases may include:


# pulling solved questions (plus some other filters?) from SUMO
# pulling solved questions (plus some other filters?) from SUMO?
# pulling a mix of solved questions (plus some other filters?) and KB articles from SUMO
# pulling a mix of solved questions (plus some other filters?) and KB articles from SUMO?
# get a kb article keyword -> article map for top 25 articles per locale and using that instead of doing lots of searches?

Revision as of 15:42, 6 March 2015

Thank you page overhaul

Summary

When a user leaves feedback using one of the forms on the site, they are redirected to a "Thank you!" page. Currently, all users see the same page regardless of their sentiment or the contents of their feedback.

We see feedback of a variety of flavors, but there are two kinds of feedback that are interesting in regards to content on the "Thank you!" page:

  1. feedback from a user who is very happy with the product
  2. feedback that's about a problem that we have known solutions for

We want to do things for both groups of people. This project covers that work broken into a series of iterations and experiments.

Status

History

  • Since the dawn of time: We've been talking about this and thinking about the kinds of things we could do and what would be most effective for users while simultaneously lamenting a complete lack of time to work on it.
  • May 8th, 2014: Will decided to put a bug on the books about this (https://bugzilla.mozilla.org/show_bug.cgi?id=1007840)
  • December 17th, 2014: Will wrote up this initial project plan.

Phase 1

This phase will focus on the following:

  1. en-US only
  2. Firefox browser product only
  3. sad feedback only
  4. require at least N words? (theory: fewer than N words will yield lousy search results from SUMO plus they're less likely asking about a problem.)

When this happens, the user will see a new "Thank you!" page which will have a set of links (how many?) from SUMO from the KB of articles that are related to the user's feedback. Each link will be instrumented. After the links, we'll have a link "This doesn't help me. I'd like to ask a question on SUMO." link.

We'll want to capture:

  1. how many people see this page?
  2. what was the feedback? what were the list of links returned from SUMO?
  3. how many people click on suggestion 1? suggestion 2? ... suggestion n?
  4. how many people click on "this doesn't help me."

We should talk to Gregg about other things to capture.

This experiment will run until we've collected enough data to warrant tweaking the experiment, committing to it ending future experiments or devising a new experiment.

Requirements:

  1. view code to figure out whether the user should see the old thank you page or the new experiment thank you page
  2. code that uses SUMO API to get back relevant links
  3. new thank you page template
  4. other infrastructure for generating the experiments
  5. infrastructure for measuring results
  6. tests for above things

Bugs:

Tracker bug: https://bugzilla.mozilla.org/show_bug.cgi?id=1007840

Bugzilla query error

Array ( [type] => error [message] => http-bad-status [params] => Array ( [0] => 406 [1] => Not Acceptable ) ) 1


Future phases

Future phases will be figured out after the most recent phase is completed so we have time to act on what we've learned.

Future phases may include:

  1. pulling solved questions (plus some other filters?) from SUMO?
  2. pulling a mix of solved questions (plus some other filters?) and KB articles from SUMO?
  3. get a kb article keyword -> article map for top 25 articles per locale and using that instead of doing lots of searches?