As a kid, I loved playing strategic warfare games like Command & Conquer or StarCraft. In those games, you usually start with a handful of resources to build a base and an army to defeat an opponent. But here’s the catch: everything around you is dark; it’s the “Fog of War”. You know your enemy is somewhere out there, but not where and what he’s doing.
Growth Marketing is similar, but instead of an opponent, you seek customer love and exponential growth. The Fog of War is the uncertainty between you and growth. You have to uncover as much fog as possible a.k.a. learning through experimentation and conversation.
Growth marketers often ask me how to regain focus (I’m a mentor in a startup accelerator). They’ve been trying to “unlock” growth for a while without finding the break-through.
I give them an analogy and three pieces of advice.
First, get back to the growth model
Getting back to your Growth Model is like getting back to the basics. And you can never do them well enough.
Setting up a Model is the first thing you should do when starting with Growth. If you haven’t done that, yet, get at it! Being overwhelmed often comes from being too deep in the weeds. Zooming out the see the big picture again helps.
The Growth Model can be expressed in ”an equation that tells you what the different variables in your business [are] and how they work together and translate into growth” (according to Hila Qu from GrowthHackers). 
The equation of a Growth Model is the product of input, which consists of three main parts: user acquisition, Aha Moment and Core Product Value. The result is output (usually Growth Rate). You want to make sure to focus on input, not output.
User Acquisition is how you get people into the door. The “Aha Moment” is the first time people understand the value of your product that leads them to retain. “Core Product Value” is the problem that your product solves on an ongoing basis. Growth Rate is the product (think: result) of everything you do to move the needle. It’s calculated by dividing this month’s with last month’s growth.
The Growth Model helps you to see all the steps in the sign-up or buyer process that you can influence.
Let’s go through a hypothetical example with Instagram.
- User-acquisition: Instagram acquired a lot of users through Facebook*
- “Aha Moment”: once they upload a picture, use a filter and get a response from someone, users understand the value.
- Core Product Value: when connecting with a certain amount of people, interacting with their pictures/videos a certain amount of times and getting the same responses for their content, users retain.
*(then Facebook acquired them lol)
Just looking at these steps, you can see how many opportunities there are to spur further growth. You could get more users or convert the ones who want to sign up better.
You might have noticed in the image above that my sketching skills are not so high I added two more steps: “invite friends” and “click ads”. You might have also noticed that the Growth Model reflects the AARRR funnel.
Note that Instagram is a “simple” app, so “Aha Moment” and Core Product value are close together. An application like Jira, with a broad range of use cases and functionality, has much more distance between the “Aha Moment” and Core Product Value.
The equation for Growth Models can have several factors for each step. There isn’t just one way to acquire users and there can be several “Aha Moments” and experiences of Core Product Value. The User Journey or Buyer’s Journey plays right into that. There are different journeys for every persona.
Good Growth Marketing is looking at the Growth Model, setting up hypotheses of how to improve every step, test hypotheses and refine them.
Coming back to the Instagram example, we could ask ourselves “would we get more users if we required only email and name for sign-ups?”. Notice how the question is not “would the Growth Rate increase if we made sign-ups easier?”. Focus on input, not output.
Second, look at the funnel
We already touched on the AARRR funnel in the previous section. The acronym stands for Acquisition, Activation, Retention, Referral, Revenue. If getting back to the Growth Model doesn’t help it’s time to crunch some hard numbers, baby!
Whereas we looked at things from a high-level in the first step, now we’re zooming back in a bit and take a more data-driven perspective. What we’re specifically looking for is how many users advance from stage to stage, and how we could get more users the final one: revenue.
Ideally, the user drop-off between each step is minimal. Here’s an example (with pretty unrealistic numbers) to make it easier to understand:
|Users (MoM)||Percentage (total)|
Looking at the drop-off at each stage can tell us a couple of things. First, note that 5% of users drop out between arriving on your site and signing up for your product (again, not realistic). But you lose 15 percentage points between the second and third stage and 55 percentage points between the fourth and fifth stage. That’s where you should get suspicious and try to find out why.
The best way to answer that question is by looking at cohorts (groups of users). In Google Analytics, you can look at cohorts by "acquisition date". But the real magic happens when you group them by actions they take or not. That’s the way you should think about this.
Let’s look at a real-life cohort analysis*.
*Note that Google Analytics is limited to cohorts by acquisition date
As you can see, there are big drop-offs between Week 1 and 2, 4 and 5, and 5 and 6 on average. The cohort from February 25th-March 3rd shows a better than average performance in the first two weeks.
To find out why we compare the well-performing cohort with the worse performing one. Then, you could look at different metrics to compare the two: where did they come from? What did they do on your site or in your product? Did they use a different device? The answer is somewhere in the data and segmentation is the fastest way to get it.
Coming back to our example, it seems that traffic from social networks seems to perform MUCH better than channels like organic search.
It doesn’t stop here, though. Once again, you should ask yourself why. Have you run a social media campaign recently? Have you changed content formats or got a lot of shares?
As I mentioned, Google Analytics has a limited functionality right now. Other notable tools are Mixpanel, Tableau, Amplitude Segment, Chartmogul, plain SQL, and tons of others.
Third, talk to your customers
Talking to your customers is one of the most powerful things to do for a business, period. I’ve never - ever - run into a situation in which talking to customers didn’t unlock a new insight. Neither in Fortune 100 companies, Silicon Valley Unicorns nor mom ‘n pop shops.
If the first two pieces of advice haven’t gotten you back on track, seeking a conversation with your users will. The first two pieces of advice were more tied to quantitative data, now we’re entering the realms of qualitative data.
It sounds easy as pie: just pick up the phone and call a customer, right? Kind of! Two questions prevail: “which customer?” and “what should you ask her?”.
The first question can be answered in two ways: you either talk to your best customers, those who love you and are already paying for your product. Or you talk to those who did not sign up (if you can), churned or are dissatisfied. Which one should you choose? Both are valuable and deceiving. Speaking to customers who love you might show you a way to get more like them. Talking to churned users can give you a wrong sense of reality because they might not have been the customers you wanted in the first place. Speaking to customers who already love you might mislead you because they already love you/the product.
I always start with the ones who signed up but aren’t happy. Also, I try to weed out the ones that weren’t a good fit in the first place. Why? Because “fixing” churn should come before growth.
How many users should you talk to? There is a gold standard in usability testing that you don’t need more than 10-15 users to get valuable feedback*. You hit a principle of diminishing returns when testing with more people. The results from 25 users won’t give more away than those from 15.
*Some argue that even 5 users are enough. 
[caption id="attachment_2455" align="aligncenter" width="500"]
(Image from Nielsen Norman Group )[/caption]
I’d argue the same holds true for viable user feedback for startups. Focus on 10-15 users who churned but make sure to qualify whether they would have been a good fit, first.
Face-to-face conversations are ideal, but Typeform/SurveyMonkey/email/insert favorite tool are fine, too. I prefer phone/email over surveys because the interaction is more personal. Customers get a different sense of appreciation.
For the second question (“what to ask?”), keep it more high-level. You should test specific questions about features or design components in a different way. Somewhere along the user journey, churned users experienced a disconnect between expectations and reality. They didn’t find what they were looking for. An explanation could be that your product doesn’t have the specific feature, overpromised, or users literally couldn’t find it.
Here are some questions to ask that will help you discover that:
- What were you hoping to get when you signed up for [product]?
- Why did you not continue with [product]?
- What would you wish [product] would do better?
- What did [product] do well?
- What would it need for you to change your mind about [product]?
When sending out surveys, keep it to 3-5 questions. Don’t forget to collect a bit of profiling data about the churned user, such as role/occupation, company, and age. Again, keep it simple. More data doesn’t always lead to more information.
“If you don’t practice you don’t deserve to win” – Andre Agassi
So, what’s the analogy I give mentees? It’s this: be a tennis player, not a golf player.
The golf player wants to get to his goal with as little hits possible. He doesn’t want to re-center or get into a rhythm. Every move is one too much. His goal is “forward”.
The tennis player is most successful when he comes back to the center of the field. He may take a couple of steps to one side, forward or backward, but if he moves too far he is vulnerable. In the center, he’s the most dangerous.
Growth Marketing is like Tennis. You should zoom in and follow the rabbit holes (think move sideways or back and forth). But make sure to always come back to the basics: getting back at the Growth Model, looking at the funnel, and talking to your users.