Who should invest in a growth-driven design website, and when is it the wrong fit?

Growth-driven design has three stages: strategy, a launchpad site, and continuous improvement. The first two look like a project, with a start and a launch. The third is designed to repeat. It runs a cycle in which the team plans around a focus metric, builds changes, learns from the results and shares what it learned with marketing, sales and service.

Lean Labs' growth-driven design program shows what that third stage asks of a client. The agency works in 30-day improvement cycles. It reviews page behavior, form results, CRM lifecycle data and feedback from the sales team together. It then ranks candidate changes by expected pipeline impact, by buyer relevance and confidence, and by effort.

For a client, that means giving the agency access to the data and to what the sales team hears. It also means naming someone who can approve changes soon enough to keep the work moving. So a good part of the fit question is about your business, and you can answer it before any design work is scoped.

The method has the most to work with when buyers use the site to evaluate you and the questions keep coming. A new product, an objection the sales team keeps hearing or a pricing change each gives you a reason to change a page and see what happens.

Some sites have one settled job. A site that lists a location, opening hours and a phone number, for a business whose customers arrive some other way, gives improvement cycles little to work on. That site still needs upkeep, such as software updates and corrections when the facts change. Upkeep is maintenance, and it can be bought without a program of experiments.

Each cycle needs a person who gathers the evidence, chooses the changes and reports what happened. That role doesn't have to be a full-time hire. It does have to belong to someone with the hours to do it, and we compare the ways to staff it in running growth-driven design in-house or with an agency.

Approval affects how quickly that person's work reaches the site. If a headline change needs sign-off from several executives and they meet once a quarter, changes wait in a queue and each lesson arrives later. One way to shorten the wait is to name a person who approves routine changes and keep group review for larger decisions. Where the approval path can't change, plan for fewer changes in a year and weigh that against the cost of continuing cycles.

The cycle's learning step needs evidence about buyers and a result you can count. Evidence about buyers comes from interviews with customers, recorded sales calls, the questions that arrive through support and what you see when people use a page. The result is a number such as demo requests or completed forms, with a record of where it stood before you changed anything.

A business with both can connect a change to a reason and a result. Record that number before and after an edit so the team has a comparison. If you're choosing what to count, we cover the options in the metrics to track in growth-driven design.

Low traffic makes numerical evidence slower to gather. A split test compares two versions of a page, and it needs enough visitors to tell a real difference from chance. Optimizely's guide to experiment length explains that the number depends on the page's current conversion rate and on the size of the difference you want to detect. A smaller difference needs more visitors. Waiting until a deadline doesn't supply them. On a quiet page, a test can run for months without settling anything.

The other ways of learning remain open. You can interview customers about how they chose you, and you can watch people use a page. Nielsen Norman Group's guidance on interviews is that they suit attitudes and memorable experiences, while usability questions are better answered by watching someone use the design. Neither method depends on how many people visit the site.

You can also make a change on judgment and say so. If the sales team keeps hearing the same question and the page doesn't answer it, adding the answer is a reasonable decision. Record it as a judgment. A before-and-after comparison is uncontrolled on any page, and on a quiet one a few inquiries either way can come from chance or from something else that changed.

Traffic is a bigger obstacle when the other sources are thin as well. A company with no customers and no sales conversations yet has less to draw on, though it can still talk to prospects and watch people use a page. We look at that situation in growth-driven design for startups.

The investment is a launch plus the cycles that follow it. If the agency budget ends on launch day, you've bought the first two stages. That can still be a good site. Improvement after that depends on someone inside the business having the time and the skills to keep working on it.

No single monthly figure marks the minimum. A monthly fee buys capacity, and the useful question is what that capacity can ship in a cycle. Knowmad's published growth-driven design tiers, for example, start at 4, 7 and 9 hours a month for $800, $1,400 and $1,800. Those are one agency's starting prices and they aren't a market rate. Where an offer is sold in hours, as these are, ask how many of the changes on your list fit into them. Other offers are scoped by deliverables or by cycle, and the same question applies in their terms. We go further into pricing in how much growth-driven design costs.

A bounded project has a defined scope, a price and an end. We'd recommend one when the need is fixed. A site for a single event, or a rebuilt site the company knows it won't staff afterward, can be bought as a finished deliverable.

It also suits a business that can't act on findings yet. If the owner's role is empty and there is no time or budget for work after launch, an improvement program is likely to produce more recommendations than changes. Under those conditions we'd buy the build alone and return to the question later.

Neither verdict is permanent. A company can name an owner, shorten its approval path or add a budget line later, and start improvement cycles on the site it already has. If you're earlier than that and still deciding whether to rebuild at all, see when it's time to invest in a website redesign.

Suppose a 30-person software company is replacing its website. Buyers use the site to evaluate the product before they talk to sales, and the sales team keeps hearing new objections. Its marketing manager can approve page changes without a committee. The sales team records its calls and shares them. The company has budgeted for monthly work after launch. Its pricing page gets few visits.

Growth-driven design fits this company. The site has recurring work to do, someone owns it, buyers are within reach and the budget continues. The quiet pricing page affects the plan and leaves the decision alone. The team should expect to improve that page through sales-call evidence and recorded judgment, and save split tests for pages with more visitors.

Now change two facts. Every page change needs sign-off from the founder, who reviews the website once a quarter. And the budget ends at launch. A finding could wait up to three months for a decision, and no budget is set aside to carry it out. For this version of the company, we'd recommend a bounded project. It should set up measurement at launch so the starting numbers exist, and revisit the question when the approval path and the budget change.

The company and its traffic are the same in both versions. Its ability to act on what the site teaches is what changed, and that is a good place to start when you assess your own business.