Omnibox - Search Queries from History - Experiment With Only Inlining to End of Word |
|||||
Issue descriptionInline autocompletion for search queries is net-positive for users (see bug 172491 comment 16) but also triggers complaints (see related bug 172491 and bug 334300). Multi-word queries tend to trigger the most complaints. Users frequently refine past searches to add, remove, or change words; I assert without data that this is more common than similar URL modifications. Therefore, aggressive past inline autocompletion is more annoying. We should try inline-autocompleting only to the end of the currently-typed word and see whether this makes the behavior feel less frustrating, and what effect it has on metrics. My suspicion is that the effect will look negative overall, but if the negative is small enough and this feels subjectively better, that could be a tradeoff worth taking. If necessary, we could try to more carefully analyze searches with common prefixes to see how often people add/remove/change suffixes, but probably just running an experiment on the above behavior change would be an easier first step.
,
Nov 30 2016
Reminder, this change affects a somewhat narrow slice of traffic (inputs for which we inline autoocomplete to something past the end of the current word). Thus, we should use "counterfactual" logging to allow analysis on just the slice of traffic for which the user could've seen a change. (Use set_field_trial_triggered() as appropriate.)
,
Jun 22 2017
I still think something here is worth doing. But none of us are planning to work on it next quarter, so it's clearly implicitly a P3.
,
Jul 19 2017
,
Jul 20
This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue. Sorry for the inconvenience if the bug really should have been left as Available. For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Jul 20
|
|||||
►
Sign in to add a comment |
|||||
Comment 1 by mpear...@chromium.org
, Nov 30 2016