The Growth Hacking Process to Supercharge Your Revenue
Every world-class growth hacker relies on a series of steps in their growth process.
We’ve broken down the six core steps below.
Define Actionable Goals
Everything begins by focusing on a narrow, actionable goal. This is important because a growth hacker can easily have a focus that is so broad it becomes meaningless. Yes, the overall goal is growth, but you don’t attain that kind of end-result without breaking it into smaller, achievable, tasks.
Let’s say you have a product and you want your DAU (daily active users) to increase, but that’s too broad of a goal. Then you decide to focus just on the retention of existing users since this will increase the DAU, but retention is still too broad. Then you decide to focus on helping current users create content because your numbers show that when someone becomes a content creator (and not just a consumer) within your product then their activity on the site is far greater. Content creation leads to retention which leads to increased DAU. Therefore, you decide to make the goal to increase content creation by 2x.
Too Broad: Increase daily active users
Appropriate: Increase content creation by 2x
Many people have a hard time knowing when they’ve narrowed their goal enough. Here is a rule of thumb that I use. Think about your goals as nested hierarchies, and until you reach the “nest” where things can actually be marked off as individual tasks which can be completed once and for all, then you’re not narrow enough. In this case our hierarchy might look something like this:
Will there ever conceivably be a day when you can mark off “grow my startup” from a to-do list? No. The goal is too broad. Is there ever going to be a time when you can say that you’ve finished “increasing DAU.” No. Too broad. However, you can mark off that you’ve “educated members about content creation through an email.” When you find yourself at the part of the hierarchy that can be checked off as done, then you’ve narrowed your goal appropriately.
Implement Analytics to Track Your Goals
Now that you’ve decided to increase content creation by 2x, the next question is, are you in a position to know if you actually attain this goal? Are the appropriate analytics in place? Here are some questions to ask yourself:
- Do you currently track content creation metrics at all?
- Do you track content creation by cohorts or just in aggregate?
- Do you track metrics around the content itself (file size, length, views, shares, etc.)
- Do you track the devices that are used to create and consume the content?
- Do you track the referring URL’s which are most responsible for content creation?
Without analytics, goals are empty. If you can’t definitively say when a goal has been reached then you have not completed the requisite requirements before moving ahead. Furthermore, analytics will give you valuable data which can change your goals. Your analytics and your goals create a reflexive equilibrium, where they actual inform, refine, and shape each other.
As an example of a reflexive equilibrium, consider this. If your goal began as “increase content creation by 2x,” but then you realize that there is something more important than just content creation in general as it relates to retention, then you might restate your goal. If content over three minutes is the only kind of content which improves the retention of the creator and the consumer then your goal might need to be refined.
One of the great things about implementing specific analytics to track goal progress is the effect this has over time on your overall analytics. Once you’ve spent a few years on a startup, attacking one goal after the other, you’ll realize that the amount of historical data you have to work with has become very powerful. Eventually, when you create a new goal you might already have the relevant metrics being tracked, and now you have past data to look at which predates even the goal creation.
All growth hackers begin with a dull axe, but the edge gets sharper as a function of time. Just don’t give up.
Leverage Your Existing Strengths
Every startup has inherent strengths or assets that can be used as leverage. When there is something at your disposal which requires little energy, but can produce big results, then you’ve found a lever.
Continuing our example from above, you may be trying to decide if you should send out the educational email first, or if you should add a “what’s new” category first. If you have 200,000 emails on file, and you have a system of email distribution that is rock solid, and you can probably create the email in question within a day, then this looks like a promising lever.
If the “what’s new” category will require at least a few days of planning, a few days of design mockup revisions, a few days of programming, and your engineers are already completely stressed out about their to-do list, then this doesn’t look like a promising route. Especially, if you’re looking for low hanging fruit.
The law of leverage essentially makes the decision for you at this point. Send out the email. Your startup’s unique leverage comes from the size of your email list and your email distribution system, not the amount of engineering horsepower that you can throw at problems. A different startup might choose a different option, but only if their leverage takes a different form.
Plans, goals, and tasks, that are stack-ranked in a vacuum, without concern for leverage, are usually misordered. Plan your attack based on strengths. I would recommend reading chapters 3-6 of this book and then give each tactic that is mentioned a score based on your company’s specific leverage.
Execute the Experiment
You’ve already completed step 1, and you defined your goal as increasing content creation by 2x. You’ve already completed step 2, and now you are tracking the necessary data that will tell you if you’re successful in your goal. You’ve already completed step 3, and now you are going to focus on educating your members through an email blast, since this is where you possess leverage. Now it’s time to execute the experiment, which means actually sending an email in this case. Here are some things to keep in mind as you execute the experiment:
1. Write Down Your Hypothesis Before You Execute an Experiment
Before you actually run the experiment you should write down your best guesses at to what will happen. Do you think this email will have a higher or lower click through rate than the emails you already send? Why do you think this? How much do you think the email will increase content creation over the next month? Will it single handedly give you the 2x content creation goal you’re shooting for, or do you think it will get you part of the way there?
It may seem silly to write these kinds of things down when you can just send the email and find out what’s going to happen. If that’s your attitude then you’re missing the point. Hypotheses are accurate reflections of your assumption before you are given the chance to rewrite the past to make yourself look like a genius.
For instance, imagine that you write down the hypothesis that the click through rate will be lower because you already send users one email a week, and you think the second email will annoy them. Then you run the experiment and it has a higher click through rate. If there wasn’t proof of your wrong hypothesis you would be tempted to rewrite history, and you would tell the team members that this is what you expected to happen because you’re such a godsend to the startup world. Hypothesis keep you honest. Now, instead of trying to prove to everyone how smart you are, the discussion is about why your assumptions were wrong. You might come to realize that you underestimate the amount your users want to be in contact with you, and this insight has benefits which stretch far beyond email.
If the idea of forming a hypothesis makes this feel too much like science and less like the traditional culture of startups, that’s probably a good thing.
2.Do Not be Naive About the Resources Needed to Run the Experiment
Anytime you run experiments it is going to disrupt the normal flow of events in your startup. First, the entire team needs to be notified about upcoming experiments so that they can be ready for any mishaps that might occur. Second, know when your startup is already resource constrained, and be mindful of this when planning your experiments. If Tuesdays are when the server is already on the brink of failure, then don’t do something that will send 30% more traffic on that day if you can avoid it. Third, if you need a certain amount of time to finish the components of the experiment before it can be ran, then don’t overlook the time requirements needed. You would do well to remember Hofstadter’s Law:
It always takes longer than you expect, even when you take into account Hofstadter’s Law.
3.Do Not Get Discouraged by the Initial Results
There is a phenomenon which I have experienced countless times, and that is the ever-present belief that whatever experiment I’m working on right now is the one that will change everything for the better. The experiment that I’m currently devoted to seems to be the obvious answer to my company’s problems. If it’s worthy of my time then it must be the thing that will allow us to reach escape velocity. Oh, the joys of the entrepreneur’s disease.
Like we mentioned in the last chapter, most things fail. It’s ok to be optimistic (hey, it keeps me going too), but then you can’t be devastated every time an experiment produces mediocre results. That’s why a defined goal should be attacked from multiple angles. Most of the attacks simply won’t work.
4. Learn from Success and Failure from Success and Failure
Data is like publicity. There is no such thing as bad publicity and there is no such thing as bad data. Even if an experiment fails you will have undoubtedly gathered a lot of information about your product and your users that can be used in future experiments. Thomas Edison failed more than 1,000 times when trying to create his light bulb. When asked about it, Edison allegedly said, “I have not failed 1,000 times. I have successfully discovered 1,000 ways to not make a light bulb.” You can learn from successes, and you can learn from failures. You only stop learning when you give up.
Optimize the Experiment
Experiments are meant to be optimized. Experiments are fluid. They are not things you do one time and then move on. You tweak experiments. You re-run experiments. You only give up on experiments when it’s appropriate to do so, not when you’ve grown tired of them.
Have a Control Group
You should always have a control group when you are able to because this will account for environmental changes that are hard to track. If you send out the email to only 80% of your users then you can track how much content creation goes up in that group as opposed to the control group. There might be an unforeseen reason, outside of your company’s control or knowledge, that has actually led to a widespread decrease in content creation on your site. Without a control group you might be led to think that the email actually decreased content creation, which would be far from the truth.
Utilize an A/B Test
A/B tests are championed by growth hackers for a reason. They’re magical! You may think you know what the subject line of an email should be to ensure it’s opened, but an A/B test will tell you the truth. You may think you know what landing page the email should send them to, so that they will start creating content, but an A/B will tell you the truth. There are very few tools that create such large gains overall.
Remember, if you are going to run A/B tests then you must decide this before you start running an experiment. Otherwise, in our example, you would have emailed everyone on your list and then there would be no one left to benefit from what you are learning.
When to Give Up on an Experiment
I usually will not give up on an experiment until my leverage has proven to be weaker than I initially thought, or I can’t logically conceive of the experiment yielding better results without an inordinate amount of resources dedicated to it.
Now it’s time to select a new experiment, or an optimized version of a previous experiment, and move through these steps all over again. If you work the system that I’ve enumerated here, then success is more a byproduct of tenacity, and less a child of luck.