It scored 38 out of 100, and the biggest single win was a phone number.
A new motorbike spare parts store with a clean look and almost nothing underneath it working. The client came asking about advertising. The honest answer was to fix the shop first.
- Client
- Sri Lankan motorbike spare parts store, named on request
- Sector
- Spare parts e-commerce, WooCommerce
- What we did
- Full UI, conversion, SEO and technical audit against a scored rubric
- Baseline
- 38/100, audited 4 June 2026
- Evidence
- Written report plus eight annotated screenshots, saved
The context
An early-stage store selling motorbike spare parts. The visual direction was fine, clean and not amateur, which is exactly what makes this case instructive. Nothing about the homepage told the owner anything was wrong. The store looked like a store.
The enquiry that started it was about advertising. That is almost always the enquiry, because ad spend is the lever people know they can pull.
The constraint
Early-stage, limited budget, and a category where trust does most of the selling. Nobody buys a brake part they cannot inspect from a shop they are not sure is real. Whatever budget existed had to go where it changed the outcome, and there was not enough of it to waste a month finding out.
What the audit found
The store had no main navigation menu. Visitors landing on a product page had no route to any other product except the back button.
It had no meta descriptions anywhere on the site. Thirty-one products carried no descriptions at all. There was a pricing bug, and a wishlist feature that did not work.
And it had no phone number.
Where the two halves connected
The temptation here is to score the site, hand over a twenty-item list, and start the ad campaign in parallel because that is the billable part. That would have been the wrong sequence and everyone would have found out slowly.
Sending paid traffic to that store meant paying for visitors who would land on one product, find no way to browse to a second, see no evidence the business was reachable by a human, and leave. The advertising would not have failed loudly. It would have produced a mediocre cost per purchase that looked like a targeting problem, and the next month would have been spent optimising audiences against a navigation bug.
Three fixes were modelled to move the score past 55 on their own: the navigation menu, the phone number, and meta descriptions across the site. None of them are advertising. All of them decide what the advertising is worth.
The result
Baseline: 38/100, from a full UI, conversion, SEO and technical audit run on 4 June 2026 against a scored rubric, with the report and eight screenshots saved. Category-level scores, the twenty-item priority list, and the three-fix shortlist all came from the same pass.
The rubric is fixed, so re-auditing the store produces a comparable number rather than a fresh opinion. The before and after are measured with the same ruler.
What we would do differently
Lead the report with the three-fix shortlist rather than the complete twenty-item table. The full list is correct and it is also where momentum goes to die. A client looking at twenty items does nothing; a client looking at three does three, and then asks for the rest.