We sped up one block of Pikapuka’s SEO report and fixed the blog layout on mobile. In the measurement on 13 July 2026, the block took 3.07 seconds instead of 21.98. This result applies to that report and the dataset tested.
Project website: Pikapuka
What needed to change
The report calculation was slow, while blog articles extended beyond mobile screen widths. These required separate checks of server calculations and page layout.
The optimisation had to preserve the final values, classification rules and dictionary priorities.
Work and results
| Work | Result | Period or source |
|---|---|---|
| How we sped up the report | 3.07 s instead of 21.98 s in the SEO report | Measured on 13 July 2026 |
| Articles and navigation on mobile | 24 checks across 6 screen sizes | 2 October 2026 |
How we sped up the report
3.07 s instead of 21.98 s in the SEO report · Measured on 13 July 2026
We examined the first SEO comparison request, before a cached result was available. Analytics took most of the time: some calculations were repeated, and PHP processed a large list of phrases.
We reused the calculated overall result when comparing search engines. Heavy processing moved to PostgreSQL, with an intermediate query result stored for reuse. Classification rules and dictionary priorities stayed the same.
We cached normalised values within the request and changed how long the existing Metrica report blocks were cached. Data or rule changes still update the cache version so reports use current values.
The old and new calculations produced matching control results. In the recorded measurement, the block’s time fell from 21.98 to 3.07 seconds.
Articles and navigation on mobile
24 checks across 6 screen sizes · 2 October 2026
In October, we removed horizontal overflow from three articles and the blog listing. We adapted pagination and preserved the text and links.
We ran 24 checks across six screen sizes. We updated the favicon, the website icon, to 120 × 120 px. These changes did not affect report calculations.
The screenshots show the current public website on desktop and mobile. The private report containing analytics counter data is not published.
How we checked the results
For the optimisation, we compared final data, classifications and search engine sections before and after the change. Timing was measured for one block using the tested dataset.
For the blog, we checked three articles, the listing and pagination at different widths. We did not submit forms or access private client data while capturing screenshots.
What to keep in mind
3.07 seconds is the execution time of one SEO report block in the recorded measurement. It does not describe homepage loading or the speed of every screen.
The blog layout was checked technically. The effect of these changes on traffic and conversion was not measured.
Before similar work
First locate the delay: server calculation or page display. For calculations, compare both timing and output; for layout, check the relevant pages at different screen sizes.