FIXED-SCOPE PACKAGE
Three templates measured, fixed, measured again.
A speed audit with numbers, a bounded fix list, and a before-and-after table you can check yourself. The audit fee is credited against the fix if you continue within 30 days.
- Audit€3503 working days, credited
- Fixfrom €9507 working days
- Templates3home, product and one you choose
- DeliverablesRaw reportsJSON, not screenshots
What is actually slow
In most small business sites the same five things repeat: an uncompressed hero image larger than the whole HTML, a render-blocking script in the head, a font loaded from a third party with no fallback, a chat widget that costs more than the page, and layout that jumps while it loads. None of them require a rebuild. All of them are measurable, and that is the order I work in.
How it runs
| Step | What you get | Price |
|---|---|---|
| 1. Audit | Baseline numbers for three templates on mobile and desktop, the ranked causes with the cost of each, and a written fix list with expected impact. If the hosting or the stack blocks the target, the audit says so before you pay for a fix. | €350 3 working days |
| 2. Fix | The agreed items from the audit list, implemented and verified on the same three templates. | from €950 7 working days |
| 3. Re-measure | The same measurement, same conditions, with a before-and-after table and the raw reports you keep. | Included within 7 days of the fix |
The audit fee is deducted from the fix if work starts within 30 days. If the audit shows the fix is not worth doing, say so and stop — that is a normal outcome and I will write it in plain language.
The acceptance test
Agreed in writing before the fix starts, measured the same way both times. Mobile first, median of three runs on a throttled profile.
| Metric | Target | If the stack blocks it |
|---|---|---|
| LCP | ≤ 2.5 s | A stated reduction, agreed in the audit |
| CLS | ≤ 0.1 | Same |
| INP (or lab TBT) | ≤ 200 ms | Same |
| Regression | No new console errors, no broken template | Not negotiable |
No score guarantee. A Lighthouse score is a diagnostic, not a business outcome, and it moves when the tool updates its formula. The numbers above are what we agree on, and they are the ones I report.
How it is measured
- Lighthouse mobile profile, three runs per template, median reported — not the best run.
- Chrome UX Report field data when the site has enough traffic to have it; otherwise the report states that field data is unavailable instead of inventing it.
- The raw JSON reports are handed over. You can re-run them yourself and compare.
- Before and after are measured on the same hosting, the same pages and the same network profile. Anything that changes between the two is named in the report.
What is not included
- Migrating hosting or CDN. If that is the real cause, the audit says so and quotes it separately.
- Removing third-party tools you want to keep — chat, analytics, pop-ups. I measure their cost and let you decide.
- Re-design, re-writing, new photography, or rebuilding the site on another platform.
- SEO content, ranking work and paid traffic.
- Ongoing monitoring after the re-measurement. A monthly watch can be added separately.
Questions people ask
Can you guarantee a perfect score?
No, and anyone who does is selling you a number that changes with the next tool update. What I guarantee is the measurement, the bounded fix list and an honest before-and-after.
My site runs on WordPress. Does that matter?
No. The method is the same; the causes differ. Plugins, page builders and unoptimized media are the usual suspects, and I will name the specific ones, not "optimize WordPress".
Do you need admin access?
The audit needs nothing but the public URL. The fix needs a repository or a staging copy; if only admin access exists, that is a risk we discuss before starting, not after.
What if the numbers do not improve?
The measured items are re-run on the same conditions. If an agreed fix does not move its metric, I keep working on that item without a new fee, or remove it from the invoice and say why in the report.
This package fixes what is measured, not everything that could be faster. The list is bounded on purpose: a fixed price needs a fixed list, and a fixed list is the only way the before-and-after means anything.