Record time to first content in histograms |
||||||||||
Issue descriptionWe'd like to determine if snippets reduce the time to first content consumption (TTFC) for our users. Approach: - TTFC starts when a new tab is opened (not necessarily the NTP, any new tab counts) - For each step in the navigation chain we start recording consumption time CT when the page is visibly rendered to the user (partial rendering counts, counting starts at 0 for each step in the navigation chain). - TTFC stops if CT >= consumption threashold N - Since we don't have a good feel for what constitutes actual consumption, yet, we record histograms for different values of N: N1 = 2s, N2 = 5s, N3 = 10s - We collect separate histograms for navigations originating from Snippets vs. other sources.
,
May 25 2016
,
May 27 2016
,
May 30 2016
nepper@: any pointers around where this definition is coming from? Is it home-grown? I'm also wondering if the priority is set correctly.
,
Jun 7 2016
Documenting our recent conversation: yes, it's home-grown. Priority: probably not required for launch, but I think it would be desirable to be able to reason about time-to-content for navigations starting with Zine vs. other sources.
,
Jun 17 2016
FF is today. Do we have updates on this? Thanks :)
,
Jun 17 2016
There's no update for this, I think it requires some more discussion. Sorry for dropping the ball.
,
Jun 17 2016
Np. I think we can do both: discuss a broader solution to this (next week e.g.), and also do a first small iteration: A histogram for TTFC (as described in the Issue description) for Snippets, i.e. we measure the time from when the NTP is opened to when the first page is rendered visibly after touching a Snippet?
,
Jun 17 2016
Talked to nepper@ and decided to put this on hold for now, inclined to mark it as 'won't fix' after a grace period. Rationale: 1. Simpler metrics are already in place (in line with #8): https://bugs.chromium.org/p/chromium/issues/detail?id=605146 2. Proposed metric involves various hard questions that would make the metric questionable and consume (compare against other usage flows), including: a. Time thresholds (absolute values, platform dependence) b. Handling of page-transition types (e.g. does clicking a link involve a user is not consuming content?)
,
Jun 20 2016
Removing milestone label and adding it to zine-mr-backlog
,
Jul 8 2016
,
Jul 8 2016
,
Aug 4 2016
,
Sep 22 2016
Grace period is over, closing bug. |
||||||||||
►
Sign in to add a comment |
||||||||||
Comment 1 by fi...@chromium.org
, May 10 2016