What is growth-driven design, and how is it different from traditional web design?

The method's own overview names three stages: strategy, a launchpad website and continuous improvement. In the strategy stage, the team studies its customers, sets objectives and collects its ideas for the site into a ranked wishlist. The launchpad is a first release built from the top of that list. The method describes the launchpad as a foundation for collecting user data and improving, and it says plainly that this release isn't the final product.

Continuous improvement is the stage that repeats. Each cycle starts with a metric the team wants to move and the ranked work it expects to move that metric. The team builds the work, studies what happened and passes what it learned to marketing, sales and service. Then the next cycle begins.

The overview models the strategy stage at 10 to 14 days, the launchpad at 60 to 90 days and each cycle at 14 days. Those numbers are the method's model, and teams adjust them to the size of the site and the people available. We cover each stage in the three stages of growth-driven design.

HubSpot announced a certification in the method in 2017. The announcement called the training tool agnostic and said a learner doesn't need to use HubSpot. We cover the history in how HubSpot defines and teaches growth-driven design.

"Traditional web design" usually means one defined project. The team agrees on the full scope, a design and a price. An agency or an in-house team builds the site and launches it, and the project closes. Good projects of this kind include customer research and testing, and many of these sites are maintained well after launch. A single project also suits real constraints, such as a budget that has to be approved once or a rebrand with a fixed date.

The difference lies in when each decision gets made. Suppose a software company is replacing a 60-page site, and its buyers decide mostly from the home page, the product pages, the pricing page and the demo request page. Suppose also that the team is unsure whether buyers want to see prices per seat or would sooner ask for a quote.

In a single project, the team settles the structure, copy and design of all 60 pages before launch. It can prototype both pricing layouts and test them with a few buyers first, and careful teams do. But the choice is still made before the page has met live traffic. Suppose the contract ends at launch. If visitors later behave differently from the test group, changing the page is new work outside the project.

In growth-driven design, the same company launches the pages its buyers decide from. The remaining pages move across with lighter changes or wait on the wishlist. The team still has to pick a pricing layout for launch. After launch, it watches how visitors use that page and hears what buyers ask on sales calls. It chooses what to build next with that evidence in hand, and revising the pricing page is already inside the plan.

So the method postpones some decisions until the team has visitor behavior to inform them. In exchange, the first release covers less ground than a full redesign would.

A growth-driven design budget covers the launch and the work that follows it. Lean Labs' published offer is one example of that arrangement. Its article on the three stages prices a launchpad, with strategy and build included, at $30,000 to $70,000 or more over 9 to 13 weeks. The same article puts a fractional team for the ongoing work at around $5,000 a month. Its growth-driven design program page describes the improvement work as 30-day sprints, with each sprint's changes picked from page behavior, form results, CRM lifecycle data and feedback from sales.

Those are one agency's figures for its own scope. They show why both costs belong in the budget from the start. The launchpad leaves some ideas on the wishlist on purpose, and the later cycles are where that deferred work and the learning happen. A company that launches and then does no further work keeps a working site, but the wishlist stays unbuilt and what visitors do goes unused. If the agency is not continuing, assign an owner for that work. We break the costs down in how much growth-driven design costs.

What the team can learn from the live site depends on its traffic and its methods. A split test shows two versions of a page to visitors over the same period and counts which one produces more of the action the team cares about. It needs enough visitors to tell the versions apart. Jakob Nielsen made a second point in a 2005 article on A/B testing. A test measures what people did under each design, and it doesn't explain why they did it.

That's why the cycles draw on more than tests. Customer interviews, watching someone try to complete a task and the questions buyers ask sales all give the team evidence about reasons, and they work on a site with modest traffic. A team with few visitors can still run the cycles this way. It should treat what it hears as grounds for a change and not as a measured result.

Following the method doesn't guarantee a cheaper site, a faster launch or more leads. It changes the order of the work, and what that order returns depends on what the team does with the cycles.

Growth-driven design suits a company that expects to keep learning about its buyers and can fund and staff the work after launch. Someone has to own the wishlist, choose each cycle's work and read the results. A single project suits a company that needs the whole site finished by a set date or has one approved budget to spend. That company can still review the live site against real visitor behavior later, as a separate decision.

If you're unsure which description fits your company, start with who should invest in growth-driven design, which covers the cases where the method is the wrong choice. If the staffing is the open question, how growth-driven design works follows one site through all three stages and shows what each one asks of the team.