Every rebuild we ship comes with two reports, mailed as PDFs, thirty and ninety days after launch. Each report is four pages. Each one answers the same four questions, with the same set of numbers, so the owner can compare the rebuild against the old site and against itself. The report exists because most redesign projects end the day the site goes live, and most redesign projects cannot say whether they worked.
Here is what we measure, why, and what we do with the result.
The four numbers
The report opens with four numbers, in this order:
- Calls per week. Tracked via a number we put on the site — a CallRail number, ported to the business's main line, with recording and voicemail. Comparable to the pre-launch baseline because we install the tracking two weeks before the rebuild ships.
- Quote submissions per week. From the form we built. Comparable because we keep the form analytics regardless of platform move.
- Cost per lead, if they run ads. Calls plus forms divided by ad spend. Often the number that moves the most, because better conversion is a multiplier on every click.
- Share of mobile traffic that reaches the booking page. A single funnel number, measured from homepage to either the phone tap or the form submit. Cleaner than overall conversion because it isolates the one path the site is designed around.
These four numbers are what we care about. Page views, sessions, bounce rate, time on page, scroll depth — all of those show up in the appendix if the owner wants them, but they are not the report.
What actually moves
In the thirty-day report, the numbers that reliably move are the first and fourth: calls per week, and mobile-booking-page reach. These tend to be up two-to-three times over the pre-launch baseline on a serious rebuild. They move fast because they are almost entirely a function of design — the sticky phone bar, the shorter quote form, the clearer hero. The site went live on Monday; by Friday, the phone is ringing more.
The second and third numbers — form submissions and cost per lead — move slower. Forms are slower because the form audience is people who are doing research, not hiring, and research windows are long. Cost per lead is slower because Google Ads needs a few weeks of data under the new landing page before it reallocates bids, and the full benefit does not land until the ad platform has had a month of signal.
By the ninety-day report, all four numbers have usually stabilized. Form submissions catch up. Cost per lead drops the way it was going to drop. Calls per week stay at the new, higher baseline.
What we do when the numbers do not move
Sometimes a number does not move. The most common case is form submissions that stay flat — the calls go up, the mobile funnel goes up, but the form stays at the level it was before.
We have a short list of things we check, in order:
- Is the form actually sending? Submission-without-delivery happens more often than it should. Bad SMTP, wrong notification email, spam folder.
- Is the form visible on mobile? Sometimes it is sitting below the fold on devices we did not test, or behind a CTA that duplicates it.
- Is there a reason the form should convert? "Contact us" has no reason. "Get a free written audit back in 5 days" has one. The form copy matters as much as the fields.
- Is the form offering something the phone does not already offer? If the phone gets you the exact same answer faster, the form is redundant. Give the form a distinct job — after-hours booking, downloadable content, a quote on a non-urgent service.
Once we have a theory, we change one thing, measure for a week, and see what happens. We do not change the whole page at once; we change the one thing most likely to be the cause, and if it does not move the number, we change the next one.
What we do not measure
The report does not include SEO rankings, because rankings move on a timeline that is not ours and because position six versus position four is often not what is costing the business leads. We report on organic traffic levels, because those are a downstream effect of rankings that we can actually see, but we do not pretend to be an SEO firm.
The report does not include "engagement" metrics. Time-on-page is a tortured number — it goes up when people cannot find what they need as often as it goes up when they are interested. Scroll depth is the same. We look at these during a tune-up if a number is refusing to move, but they are not part of the headline.
The report does not include compliments. People sometimes email the business and say the new site is nice, or that it looks professional. The owner will tell us, and that is nice to hear, but it is not evidence of anything. A site that looks professional and does not generate leads is not a successful rebuild.
The ninety-day decision
At ninety days, we have enough data for the owner to make a decision about what to do next. The common paths:
- The numbers moved. Keep the site as-is. Most rebuilds land here. We stay on for a quarterly audit at a flat fee if the owner wants it; most do not need it for another year.
- The numbers moved, but not enough. Usually means the site is working but the traffic is the bottleneck. We refer to an SEO or ad partner we trust — we do not do paid media ourselves — and the site is ready to convert whatever they send it.
- The numbers did not move. Rare, but it happens. When it does, we do a second free audit, out of our pocket, and figure out whether it is the site or the business. Sometimes it is the business: the service offering changed, the market softened, the owner is off the phone too often. Sometimes it is the site: we got the hero wrong, the form is still too long, the target customer is not who we thought. Either way, we own the answer.
The case for writing it down
None of this is novel. Everyone who ships websites for a living knows which numbers matter. What is novel — or at least uncommon — is writing it down, mailing it to the owner twice, and being accountable to it.
A rebuild is a meaningful amount of money for a small business. Four-figure engagements for us are five-figure decisions for them, once you factor in the hours they are going to spend on content, review, and launch. The report is how we stay on the hook. If we got the bet wrong, the report is what makes it impossible to pretend otherwise.