The verdict in three sentences
On a B2B catalogue of 50,000 SKUs, an INP of 450 ms means a repair shop buyer clicking a "make, model, engine" filter waits almost half a second before anything happens, and they leave sooner than you think. The cause is almost always product lists and filters rendered in the browser, rarely the hosting. A focused audit followed by a rebuild of those two components brings INP below 200 ms in about six weeks, with conversion gains around 12% seen on comparable sites.
Reading the numbers: what INP and LCP actually measure
In March 2024 Google replaced FID with INP (Interaction to Next Paint), which measures the latency of every interaction in a visit, not just the first one. For a distributor whose customers chain OEM part number searches, filters and add to cart, it is the metric closest to the real experience. LCP measures when the largest visible element renders, usually the product image or search banner.
| Metric | Distributor site (mobile, field data) | Google "good" threshold | Likely cause |
|---|---|---|---|
| INP | 450 ms | under 200 ms | filters recomputed in JavaScript over 50,000 rows |
| LCP | 4.2 s | under 2.5 s | unsized images, render-blocking font |
| CLS | 0.18 | under 0.1 | promo banner injected after load |
| TTFB | 1.1 s | under 0.8 s | uncached catalogue queries |
| Initial JS weight | 1.9 MB | under 400 KB advised | full filter library loaded on every page |
| Share of "poor" URLs in Search Console | 68% | 0% | list and product templates |
Lab data (Lighthouse) helps with diagnosis, but only CrUX field data feeds rankings. On a B2B site with modest traffic, data is often aggregated at origin level, which is why we install in-house RUM tracking during the audit.
Budget and scope: audit, list rebuild, filters
Cost depends mainly on how the catalogue is served. These are 2026 ballpark figures for the US market on a site of this size.
| Workstream | Content | 2026 ballpark (USD) | Duration |
|---|---|---|---|
| Performance audit | CrUX, Chrome traces, filter profiling, prioritized plan | 4,000 to 8,000 | 1 to 2 weeks |
| Product list rebuild | server rendering, pagination, virtualized long lists | 5,000 to 10,000 | 2 weeks |
| Filter rebuild | server-side facets or dedicated search engine | 5,000 to 9,000 | 2 weeks |
| Images and fonts | AVIF/WebP, explicit sizes, LCP preload | 1,500 to 3,000 | 3 to 5 days |
| Search engine (Meilisearch, Algolia) | indexing 50,000 SKUs and OEM cross-references | 0 to 700 per month depending on plan | in parallel |
| RUM tracking and alerts | INP dashboard per template | 1,000 to 2,000 | 3 days |
In total, plan for 15,000 to 30,000 USD for the audit and the list and filter rebuild, over 6 weeks. The most profitable levers are nearly always the same: move facet computation out of the browser, split JavaScript per template, and stop rehydrating 200 product cards on every click.
The six-week plan
- Weeks 1 and 2: audit, field measurement, identify the three slowest interactions (usually vehicle filter, price sort, add to cart).
- Weeks 3 and 4: rebuild lists and filters, plug part number search into a dedicated index.
- Week 5: images, fonts, CLS removal, category page caching.
- Week 6: progressive rollout, RUM monitoring, confirm gains in Search Console over a rolling 28 days.
Watch out for SEO. Filter pages need clean URLs and a clear indexing policy (canonical tags, noindex on useless combinations), otherwise the performance win is paid for with duplicate pages.
Mini case study
Need a professional website?
Kolonell builds websites that attract clients, optimized for the Sénégalese market. Free quote in 2 minutes.
Daniel, CIO of an auto parts distributor in Miami, serves 400 repair shops ordering online. The site handles 9,000 orders a year at an average of 350 USD, so 3,150,000 USD of online revenue. He commissions a 6,000 USD audit and a 16,000 USD list and filter rebuild, 22,000 USD in total.
After rollout, mobile INP drops from 450 to 170 ms and LCP from 4.2 to 2.1 s. With a 12% conversion lift, online revenue grows by about 378,000 USD a year. At a 25% gross margin, that is 94,500 USD of extra margin: the investment pays back in under 3 months. Even with half the lift, payback stays under 6 months.
FAQ
Will better hosting fix a 450 ms INP?
Rarely. INP measures browser work after a click, so a faster server improves TTFB but not filter recomputation. In audits like this, over 70% of the gain comes from front-end code.
How long before Google reflects the change?
CrUX data uses a rolling 28-day window. Expect 4 to 6 weeks after rollout before URLs turn "good" in Search Console.
Do we need to change e-commerce platform?
Not necessarily. For most catalogues of 20,000 to 100,000 SKUs, a targeted list and filter rebuild in the 10,000 to 19,000 USD range is enough, while a full replatforming often runs 60,000 USD or more.
Is the 12% conversion lift guaranteed?
No, it is a ballpark seen on comparable B2B sites. It depends on your baseline, mobile traffic share and catalogue quality; a 4-week A/B test measures it properly.
What happens to indexed filter pages?
Keep combinations with real search demand (for example "brake pads Ford F-150 2019"), often 300 to 1,000 URLs, and set the rest to noindex to avoid duplicate content.
Let's scope your project. A Core Web Vitals audit of your catalogue, a rebuild of lists and filters, and a production rollout in roughly six weeks. Detailed quote within 48 h. WhatsApp +221 77 596 93 33.
Mohamed Bah
Fondateur, Kolonell
Passionate about digital and entrepreneurship in Africa, Mohamed has been helping Sénégalese businesses with their digital transformation since 2020. Founder of Kolonell, he believes every SME deserves a professional and accessible online présence.
