What do real growth-driven design examples and case studies teach us?

Lean Labs' account of the Integrate redesign describes work on the message as well as on the pages. The team clarified Integrate's value proposition and tested messages and calls to action. It also simplified the main menu and offered visitors a preview of the platform.

A layout arranges whatever message it is given, so if the team hasn't settled what the company is claiming and for whom, a new design rearranges an unclear claim. Treating the message as its own task gives the design something definite to arrange.

Within 60 days of launch, the agency reports that the rate at which visits became contacts rose 271%. It gives no starting rate, visit counts or contact counts, so the account shows a proportional change and not how many more contacts the site produced. Total leads rose 174% against the same period before launch, and an eBook co-published at the time brought in some of them. Excluding the leads attributed to that eBook, the agency reports a 45% increase.

The 271% is a change in a rate, and the two lead figures are changes in a count. So the 45% doesn't revise the conversion figure, and it doesn't remove every effect the eBook had on the site's numbers.

Because the message, the menu and the preview offer changed together, the account can't say how much each one contributed. It also gives no reason to think every site should offer a preview or shorten its menu.

Media Junction's account of the Pittsburgh Parks Conservancy site, dated October 1, 2016, is most useful for its small changes. One was a single word. The menu label "Support" became "Give," and the team made it more prominent.

After launch, on one page the team read visitor behavior as a sign that people thought the page ended at its cards. It moved the card text, and it reports that visitors scrolled further afterward. That explanation was the team's interpretation. More scrolling fits it, but an uncontrolled before-and-after comparison can't confirm it. The team also revised the summer-camp page layout and a video's file size, and the agency describes the work as continuing on a retainer.

Each of these changes began with something specific that someone noticed, and each was small enough to make and then watch. That is a fair picture of the improvement stage of growth-driven design.

For results, the account reports a 32.4% month-to-month increase in conversions during the first quarter after launch. That is a rate of monthly growth. It isn't a total for the quarter or a figure for donations, and the account gives no conversion counts to set it against.

A reported gain becomes useful when you know the metric, the period, the starting point and what else was happening.

Suppose two sites each report a 271% rise in conversion rate. One went from 0.5% to about 1.9%. The other went from 2% to about 7.4%. The headline is identical, and the two sites are in different positions. On a page with few visits and a low starting rate, a handful of extra contacts can produce a large percentage.

The period matters because the two cases don't use the same one. One figure covers a fixed window after launch, and the other is growth from month to month. You can't set one beside the other and say which site did better.

Concurrent work is easy to leave out, and the Integrate account is worth reading because it names some. A launch can coincide with a campaign, a season or a new piece of content, and each can move the same number in the same weeks. Both cases compare a site with its own past, so neither can separate the redesign work from those other influences. And an agency chooses which projects to write up, so two good outcomes don't tell you what a typical one looks like. We look at the published figures for the method as a whole in growth-driven design statistics.

The practices carry over better than the percentages. Settle the message with evidence from buyers before you commit to a layout. Lean Labs' current growth-driven design program describes a similar habit: it reviews page behavior, form results, CRM lifecycle data and sales feedback together before it ranks changes. We cover the research behind that step in how Jobs to Be Done and user research shape a strategy.

After launch, keep looking for small, specific problems. Write down what you saw and what you think it means before you change anything, so the interpretation is on record as a guess. That is the start of a hypothesis you can test.

Then keep the facts a reader of these cases has to go without. For each change, note the metric, the count you started from, the dates you compared and anything else that launched in those weeks. After a few cycles you have a history of your own site to plan from, with the starting points on record.