One way to make client sites fast, whatever they were built on

The portfolio is the problem, not the site

Speeding up one site is a solved problem. There is a plugin, a guide, a checklist, and someone on the team who has done it before.

What does not survive contact with an agency is the second site. It runs on something else. The plugin does not exist there, the guide assumes a CMS you do not have, and the person who did it last time is now the only person who can do it again. Multiply that by a client list and speed stops being a technical task and becomes a scheduling one.

That is the specific thing a delivery layer changes. It works on the page as delivered, so what it does is the same regardless of what produced the page — and the work of adding a client is a DNS record rather than a project.

What we measured

The figures above come from our own audit base, and the sample size and date are stated with them. The one that matters here is the number of different platforms in it: the same pipeline, unchanged, across every one of them.

The number of sites where the mobile score went down is printed beside it, because that is the question you should be asking on behalf of a client rather than the one a vendor volunteers. Alongside it, the count of sites that started in Google’s “Poor” band and left it — the single most reportable outcome in the set, because it is the band where a client’s competitors are ranked above them for reasons the client can actually fix.

If it is just you

A one-person studio has the same problem as a twenty-person agency and less room to absorb it. You cannot keep a specialism in six platforms current, and the hour you spend re-learning where the render-blocking CSS lives on a builder you touch twice a year is an hour nobody pays for.

The part that changes for you specifically is the reporting. A client who asks whether the work was worth it gets a page with their own numbers on it, produced by a third party rather than by the person invoicing them — which is a different kind of answer than a screenshot you made. There is no minimum number of sites: Fast covers one, and Growth only becomes cheaper than several Fast plans once you actually have several.

Where this is not the answer yet

One place, and it is better you hear it from us.

Stores. Five is not a sample. We can run the pipeline on a shop as easily as anything else, but we cannot show you a number that means anything about ecommerce, and the small one we have runs below the rest of the base rather than above it. If shops are most of your list, treat this as unproven for your case and check one of them for free instead of taking our word for it.

What to do with this

Run the free check on the client site you are least happy with. It takes about a minute, it produces the same before/after page you would send them, and it costs nothing to find out that we are not the answer for that site — which happens, and the report says so when it does.