How should high-growth companies evaluate ROI-driven website design services?

Before the first sales call, pull four figures from your analytics and CRM:

• visits to the pages that lead to a sale
• leads those pages produce
• the share of those leads that become customers
• what a new customer contributes after the costs of serving them

A fast-growing company has a particular reason to do this first. Suppose leads from your site have been rising 4% a month, and assume for illustration that the rate continues. After twelve months they would be about 60% higher with no redesign. Nobody can be sure the rate will hold. Reading the whole before-and-after increase as the redesign's effect would leave that earlier trend unaccounted for. Write down what you expect to happen without the project, so there is something to compare the result against.

Keep the four figures separate when you talk to an agency, because a website affects them differently. A redesign works most directly on the share of visitors who become leads. Traffic depends on marketing as well, and the share of leads who become customers depends on your sales team too. When an agency names an expected improvement, ask which of these figures it means. We list the measures worth recording in what metrics to track in growth-driven design.

A return has a cost side. Count the full cost once: the build, any continuing work after launch, and the hours your own team will spend. A quote may already include some of these, so check before adding them. Suppose a company has the following figures.

The first-year cost is covered when the project brings in 12 customers who wouldn't have come otherwise. That figure is $108,000 divided by $9,000, and it counts each customer's first year of contribution.

The timing needs a note. For a customer who signs in month ten of the project year, the first year of contribution extends into the following year. So the calculation tells you whether the customers won are worth the cost. It doesn't tell you the project has returned its cash by the end of month twelve.

Notice also that the calculation uses contribution and leaves revenue aside. Twelve customers at $15,000 is $180,000 of revenue, which looks well ahead of a $108,000 cost. After the cost of delivering the product and winning each customer, those twelve return $108,000.

To see how large a change that is, set the growth example aside and make a simpler assumption. Suppose the site would have kept producing 40 leads a month without the project, with one in ten becoming a customer. That is four customers a month. Assume the number of new customers stays flat at four a month through the nine months being compared. If the build takes three months, the new site is live for nine months of the year, and the old site would have produced 36 customers in that time. Twelve extra customers is a one-third increase.

Now ask each agency what change it expects and why. Its answer is an assumption until it happens. Suppose an agency expects the new site to produce 10% more leads from the same traffic. With the same close rate, that is 10% more customers, against the one third this company needs. The gap is visible before anyone signs.

A portfolio shows that an agency can produce a good-looking site. Unless it comes with results, it doesn't show what happened after launch. Lean Labs makes this point in its guide to choosing a growth-driven design agency. The guide advises buyers to ask for before-and-after numbers on specific pages, such as bounce rate, exit rate and conversion rate, and to be wary of a pitch that offers the portfolio alone.

That is the right first question. A before-and-after number is still an observation about another company, so it needs a few more:

• What was the starting figure, and over what period was the change measured?
• What else changed in that time, such as ad spend, pricing or a product launch?
• Is the result counted in leads, in pipeline or in closed revenue?
• How was the result credited to the website?

Ask the agency to answer these from its project records. The answers tell you how much weight a case can bear. What you're looking for is evidence that separates the website's effect from the other things that were happening. We look at how to read published examples in what growth-driven design case studies teach.

When an ROI claim rests on an attribution report, it helps to know what the report does. HubSpot's attribution reporting measures three different conversions: contacts created, deals created and deal revenue. The deal and revenue reports are available only in Marketing Hub Enterprise accounts, so ask which report a revenue figure came from. For each report you choose a model. First touch gives all the credit to the first interaction, and last touch gives it to the final one. Linear spreads it evenly, and other models weight interactions by timing or by how often each type occurs.

The same closed deals can credit the website with different amounts depending on the model you pick. That shows what attribution is. It divides credit among the interactions the system recorded. It doesn't test whether a deal would have closed without the new site.

A controlled comparison gets closer to cause. For a single page, HubSpot's A/B test shows two versions to equal halves of the visitors over the same period, so outside events affect both. If a controlled rollout isn't available for the whole relaunch, compare the results with the expectation you wrote down beforehand and record everything else that changed. That comparison is an observation, and it can't separate causes the way a test can. Agree on the method with the agency before the work starts, so nobody chooses the measure after seeing the results.

Results depend on the work being delivered, so look at how it's scoped. Lean Labs' guide suggests asking for a phase-by-phase plan that says what you receive and when. It also suggests asking what stays with you if the engagement ends early. Check that your calculation counts the continuing costs once, including the platform subscription and any monthly retainer. Ask as well whether your team can edit pages itself once the site is live, because developer time for routine changes is another cost. We cover the terms in what a growth-driven design contract and retainer should include.

Some agencies let you test the relationship before the full commitment. Lean Labs sells its Design Blueprint as a separate four-week project ahead of a full build. The price starts at $6,000, and a client who is unhappy in the first three weeks gets the fee back. A step like this limits what you risk while you find out how the agency works and what direction the design will take. It isn't evidence about revenue, and it doesn't replace the calculation above.

If the expected change falls short of what your calculation requires, you have an option before saying no. Ask for a narrower scope, such as the few pages that lead to a sale, and have it quoted separately. Then run the calculation again with the new cost and the change the agency expects from the smaller project. The answer may improve or it may not, and the figures will show which. We explain when a rebuild is the wrong tool in whether a website redesign will fix low conversion rates.

Fast growth also changes the inputs while you're deciding. Contract values, close rates and traffic can all move within a quarter. Rerun the calculation with current figures before you sign, and again when the first results arrive.