Why startups should begin by doing things that don't scale, and how to know when to switch to things that do

• Before startups commit time and resources to things that scale, the story of a popular urban-gardening startup in London shows why they should focus first on things that don't.
• Franky Athill, head of marketing for Patch Plants, said his early team of five got its first sales by targeting outreach efforts exclusively on a few apartment buildings.
• The team grew and personally delivered plants to the small pool of customers.
• Once the idea gained traction, the team shifted gears to rapidly ramp up daily sales from 10 to 1,000 in three years and expand into Europe.
• This article is part of a series on growing a small business, called "From 1 to 100."

Focusing on how to grow before you have enough data to inform your decisions can be a waste of scarce resources, according to Franky Athill, head of marketing for the popular UK urban-gardening startup Patch Plants.

"In the first one or two years in business, do things that don't scale," he said in an interview with Business Insider.

Until your business begins generating large enough data sets to support rigorous analyses, focus on your team's anecdotal assessments about what is and isn't working for your business.

"If you don't have enough data to actually get statistically significant results, don't look at the data at all," Athill said. "Just look at qualitative data and make your decision based on experience, advice, and intuition,"

Here's how to tell whether your startup is ready to shift into growth mode.

Start with an extremely specific focus

Before Amazon sold, well, everything, it was an online bookstore; before Facebook reached billions of people around the world, it connected students at a handful of colleges; and before Patch became an international business, it served a few apartment buildings in London.

Founded by Freddie Blackett in 2015, Patch is a direct-to-consumer service that helps would-be green-thumbs discover the right houseplants for their homes or offices, sells and delivers them, and provides tips for how to keep the plants alive.

When Athill joined as the fourth member in 2017, the Patch team not only grew and delivered every plant it sold, it would also raise brand awareness by ringing buzzers and slipping fliers under every door in selected buildings.

"We made it super, super specific to that building, where the flyers actually had a picture of the view from that building," Athill said. "So it had a huge conversion rate because it was extremely relevant to the people who got the flyer."

The team understood that doing everything themselves was not a viable long-term approach, but that hyper-focus on establishing their business' value proposition got them pointed in the right direction.

"We would never consider doing anything like that now, but it was an excellent way to get going," Athill said.

Information from those early sales provided the foundational dataset that they could then analyze to develop a more scalable business, though Athill admits they didn't quite get it right at first.

"We've moved from a company that did data badly to one that does data excellently," he said.

Metrics without targets are meaningless

Even if your startup is not yet ready to draw meaningful conclusions from its initial metrics, Athill said one of the first-priority hires should be someone with the analytics skills to digest and communicate those numbers for the team.

Once there is enough data to identify statistically significant trends (numbers like 1,000 web sessions or 10 sales are not enough, Athill said), the next task is to set specific targets for each role at the company.

"We made a huge mistake for a long time where we would report on trends, but that's really useless because you can't actually take any action against that," he said. "You have to set a target line."

Your data person should also be able to meaningfully integrate several different information sources into a user-friendly dashboard of relevant key performance indicators for your business.

Athill said he sees a lot of startups "making decisions based on rough or inaccurate data, or they're having to go to three different sources to look at three different views on the same kind of metric. And none of them are that accurate."

By focusing Patch's early efforts on generating lots of data, and later efforts on actually relying on it, the company has been able to design a marketing and distribution strategy that is truly scalable.

Far from individually growing and delivering plants in London, Patch now has a fulfillment hub in Holland that will enable it to reach most of Europe, and a pilot operation in Paris.

In order to get big, it seems, startups must first think small.