Growth Doesn’t Start With Experiments.

It Starts With Understanding Relationships.

Spend an hour on Growth Hackers and you’ll find lots and lots of comments, tactics, and studies about companies that grow. Take HubSpot, for example:

“HubSpot uses webinars to educate users, generate buzz, and attract thousands of new website visitors — a number of whom will eventually become customers.” [Source]

Webinars worked for HubSpot, so they should work for you too, right? Webinars work.

It’s easy to draw ideas from a bucket of things that other companies did. Unfortunately for you, other companies are not your company. Their teams made decisions in the context of their company and not yours. You’ll quickly understand this as you watch your exact reproductions return mediocre results.

Growth doesn’t come from reproducing the experiments of successful companies, it comes from understanding relationships.

In my experience, my most impactful learnings emerged from testing the relationships between a segment, an action, and a theme. Those learnings turned into bets the larger business could run with. Once you begin to frame questions in this context, every experiment you run and all research you conduct will have targeted business impact.

Segment

A group of customers or users who share a similar set of properties. One segment could be customers who work in construction. Another segment could be customers who have paid you more than $60. If you desire, you can add more traits to the list to create an even more focused segment (like customers who have been paying for more than 6 months but not more than 12 months and who show lower-than-average usage and haven’t talked to the support team).

Action

A simple, measurable behavior. This can be something like requesting a demo from the sales team or perhaps signing up with a discount code. Actions are often easier to track if you build a software product because you can get very granular with things like exploring additional pricing details or creating their first marketing campaign.

Theme

This is the one that people often struggle with. A theme is a thing that you think is true about your customers in a general sense. Usually, people forget to generalize and they skip right to a specific example of the theme. Below are some examples of the difference.

The power of these three things comes when they are combined into a single relationship statement:

“How can we test the influence of [THEME] on how often people [ACTION] for those people who [SEGMENT]?”

Starting to make sense? How about if we fill in the blanks:

“How can we test the influence of social pressure on how often people request a demo for those people who are first-time visitors?”

At this point, magic happens. There is no limit on what research you do or what testing tactics you use. You can now excuse experiment failures without making assumptions about why they failed. You are beginning to draw a map of relationships that can be used now and in future projects.

Themes — where do they come from?

As I mentioned above, themes are the most challenging part of this framework. They are a bit abstract and force you to think in terms of commonalities instead of specific details. Even I face challenges wrapping my head around themes from time to time.

In most businesses, there is a need to increase the frequency of certain things that customers do, whether it’s paying money, leaving reviews, asking for help, creating content, or simply returning to your store to look around. Since you already have an action and some form of a segment, all that’s missing is a theme. So, where do themes live?

A very common mistake is to ideate the themes from nowhere. Usually, a team will brainstorm for hours and generate a list of clever ideas based on bias or past experience. You’ve probably heard this a thousand times, but there is a better way, and that’s to get out of the building and talk to customers.

Start talking to people from each segment who didn’t do something.

Ask them why they didn’t do that action. Ask them about their experience up to that point. Ask them a whole bunch of other stuff if you’re curious (but stay organized and specific).

Once you have all of your answers, group them by generalized stories:

On the data analytics side (if you have that stuff), you can compile a list of things that users did do instead. This might give the user statements some weight or perhaps reveal strange patterns of use.

Now you have a list of things people said validated by a list of things people did. Together, these make up themes.

Once you have all three, start brainstorming

Pull together a group of coworkers who build for or talk with customers (including but not limited to engineering, marketing, sales, services, or support) and present them with a prompt:

Sometimes there is only one way to test this. Sometimes there are hundreds of ways. You may find yourself getting through 5 or 10 different prompts during the brainstorm, but you’ll end up with a hyper-focused list of experiment and research ideas.

(Want a more granular example of how this works from beginning to end? Check out the expert’s document.)

Now you’re asking the right questions

One more time so you don’t forget this part:

Your goal is to find ways to test the relationship between a segment, an action, and a theme.

Never again will you copy and paste the work of another successful company without the consideration of your funnels. All of your ideas are now coming from the reality of what your customers do and feel, backed up by the truths that you pulled from past research and experimentation.

Sometime in the future, when you have spent hours doing your best work, you’ll find an email in your inbox from the CEO with a chart attached.

That chart goes up and to the right.

Learned something? Click the 💚 to say “thanks” and help others find this article.

Geoff Daigle is on the growth team at HubSpot. You can find him on twitter at @dailydaigle