FY21-Q4 Verify Frontend Key Result (Product): Reduce the average LCP on verify pages to <4s `=> 68%`

The verify pages rank quite poorly on our Sitespeed dashboards. Especially when focusing primarily on LCP (Largest Contentful Paint). Where an LCP of <2.5s is good, >2.5s && <4s needs improvement and >4s is poor

If we filter by only the verify pages (Dashboard) we see that the baseline for this is that 30% of our tested pages have a "good" LCP. Some of our worst performers have an LCP of >10s, with an overall average of 6.13s (in October).

To measure this I want to get the average out of the poor category and into needs improvement. Tackling the poorer performing pages will have the biggest impact here.

Progress

100 * (baseline - current) / (baseline / target) = completion %

100 * (6.13 - 4.678) / (6.13 - 4) = 68.169%

Month	Average LCP
October (Baseline)	6.13s +0 Link
November	6.06 s -0.07 Link
December	5.015 s -1.11 Link
January	4.68s -1.45 Link

---

Retro

This KR went better than expected. The Verify pipeline list is the only holdout and we have an issue scheduled to take care of that. If we remove those pages from the equation, we end out with and average of 2.978s which would score us a 158% completion (OKR scoring is hard).

Stop

A good portion of this effort ended out being backend work. I should have communicated that a little more effectively as it seems to have been perceived as a frontend only effort. If I'd spoke up about the backend effort sooner, that scheduled issue may have been a completed issue by now.

Start

Having dashboards for this was great, but I would like them to be more visible. If we have grafana on a KR in future, I can embed the result directly like this:

Edit: That doesn't actually work 🤦

Continue

Alongside the tracking, I included a but of history about the why and not just the what on this KR. I think that really helped.