Before and After Means Nothing Without a Baseline
Every website rebuild ends with the same slide. The old site on the left looking tired, the new one on the right looking sharp, and an implied round of applause.
It is a satisfying slide and it proves nothing. Prettier is an opinion. You paid for better, and better is a measurement.
The baseline is the whole game
Measure the old site properly before it disappears. Not from memory, not from a report somebody ran last year with different tools, and not reconstructed afterwards from whatever happens to be in an archive. The day the new site goes live the old one stops being measurable, and any improvement claimed after that is storytelling.
Make the baseline broad, because a rebuild changes more than speed. Record how fast the pages load on a phone. Record what the pages actually say: the titles, the headings, the prices, what somebody is asked to do. Record how many enquiries and calls came in per month for the past year, and which pages people arrived on. Most of this takes an afternoon and it can never be collected later.
Measure both sides with the same instrument
Different tools give different numbers for the same website, so a before taken with one tool and an after taken with another is a comparison of tools not of sites. Use the same test, on the same kind of connection, on the same pages.
Keep the output. A screenshot of the before, saved somewhere that is not the laptop of whoever is building the new site, is worth a great deal at the point where everybody is arguing about whether it is faster.
Judge against a yardstick you did not invent
Numbers persuade when the target was set by somebody else. Core Web Vitals are published thresholds for how quickly a page becomes usable and how much it jumps around while it loads. Neither you nor the person building your site controls them, which is exactly what makes them useful.
Put the targets in the brief before the build starts. A builder who knows the standard will be applied builds to it. A builder who first hears about it at the comparison is entitled to feel ambushed, and will usually be right.
Report movements, not adjectives
An honest report is a short table. This measurement went from here to here against this threshold. These pages now state a price where they did not. This form went from nine fields to four. Enquiries by month look like this.
Where something got worse, and on a real rebuild something usually does, write that down too, with what is being done about it and by when. A spotless report reads as curated, and most people can feel it.
Be careful about what you claim caused what
A rebuild changes several things at once, which makes attribution harder than anybody wants it to be. A guesthouse that relaunches in October and has a strong December has not proved anything about its website. A practice that starts a radio campaign in the same month as a relaunch has made the question unanswerable.
Compare like periods where you can, note what else changed, and treat the direct measurements, such as speed and what the pages say, as the firm evidence. The commercial numbers are worth watching and worth being honest about. You are allowed to say that enquiries are up and that you cannot yet separate the website from the season.
The fortnight after launch is part of the measurement
A rebuild usually changes the web addresses of your pages, and the old addresses are sitting in Google, in printed material, and in the bookmarks of people who already know you. If nobody redirected them, those visitors now land on an error page, and the first thing your shiny new site does is lose the customers who were already looking for you.
Check it yourself in the first fortnight. Search for your business, click the results, and see where they land. Ask the builder for the list of redirects and spot-check five of them. This is also the window where a form that worked in testing quietly stops delivering, so send yourself a test enquiry on day one, day three and day fourteen.
The brief is where this is won or lost
Trace a disappointing rebuild backwards and it almost always starts in a brief that specified the appearance and nothing else. If the brief never committed to a speed target, nobody built for one. If the brief never said what each page should make possible, nobody built that either.
A good brief says what the site must do, what it must never do, and what will be measured afterwards. It also says that nothing may be invented to fill a gap. A page of plausible-sounding text about your business, written by somebody who has never visited it, is worse than an obvious blank, because the blank gets filled in and the invention gets published.
The takeaway
Before you sign off on any rebuild, capture the baseline and store it somewhere that does not belong to the person doing the work. An afternoon of measuring makes every claim about the project checkable for years.
Amysa measures the site it is replacing before it replaces it, and puts the targets in the brief not in the compliments afterwards. Before and after is a reasonable thing to show a business owner. It just has to mean something.
KEEP READING
Related field notes
Tell us what still happens on the phone.
We will tell you what the site could take off your hands and what that costs. The conversation is free.