There is a revolution taking place in the world of startup growth, and we wanted to help people understand this new phenomenon.
Those who understand growth hacking will have a competitive advantage that is hard to overstate, and we wanted to provide a robust framework for thinking about it.
This guide is for entrepreneurs, founders, growth leads, or anyone else who is trying to grow a startup.
If acquiring new customers (and retaining existing ones) is important to your business then you should read this guide.
If customers matter to you, then growth hacking should matter to you. Let’s begin.
What is Growth Hacking?
Growth hacking is so misunderstood that there is a desperate need for this post. Few concepts have been as polarizing and revolutionary, simultaneously. Is it marketing in disguise? Is it a buzz phrase used to increase salaries? Is it the future of internet products?
Let’s start at the beginning…
The short history of a controversial concept
The phrase “growth hacker” was coined by Sean Ellis in 2010. When I asked Sean why he felt the need to coin a new phrase he said that it stemmed from his frustration when hiring replacements for himself. I’ll explain.
Sean had helped a number of internet companies achieve incredible growth, and a few of them even had an IPO. Needless to say, Sean became the guy that the valley went to when they needed to grow their user base, and he would take equity and payment in exchange for his services. He essentially became a one man growth shop, setting up systems, processes, and mindsets, that could be maintained after he left. Eventually, he would hand over the keys to his growth machine to someone else, and he would ride off into the sunset. This is where the problems started.
While searching for his replacement he would often receive resumes that were legit, but not relevant. They had marketing degrees, and they had marketing experience, but they were still missing something. Sean knew that the kind of strategies he employed did not represent the typical playbook used by traditional marketers, and if he gave them the reins it would not be a good fit.
A traditional marketer has a very broad focus, and while their skill set is extremely valuable, it is not as necessary early in a startups life. In the first phase of a startup you don’t need someone to “build and manage a marketing team” or “manage outside vendors” or even “establish a strategic marketing plan to achieve corporate objectives” or many of the other things that marketers are tasked with doing. Early in a startup you need one thing. Growth.
Sean asked for marketers. He got marketers. So Sean changed what he asked for. The title of his watershed blog post was “Find a Growth Hacker for Your Startup” and the idea was born.
A growth hacker is not a replacement for a marketer. A growth hacker is not better than marketer. A growth hacker is just different than a marketer. To use the most succinct definition from Sean’s post, “A growth hacker is a person whose true north is growth.”
Every decision that a growth hacker makes is informed by growth. Every strategy, every tactic, and every initiative, is attempted in the hopes of growing. Growth is the sun that a growth hacker revolves around. Of course, traditional marketers care about growth too, but not to the same extent. Remember, the power of a growth hacker is in their obsessive focus on a singular goal. By ignoring almost everything, they can achieve the one task that matters most early on.
This absolute focus on growth has given rise to a number of methods, tools, and best practices, that simply didn’t exist in the traditional marketing repertoire, and as time passes the chasm between the two discipline deepens.
Traditional marketers are skilled at understanding traditional products, but the internet has created a radical redefinition of the word product. For thousands of years a product has been a physical good, but now they are invisible bits and bytes in the form of software products. Products used to only be things like cars, shampoo, couches, and guns. Now Twitter is a product. Your online accounting software is a product. Things you can’t hold, per se, are products. This transition is most responsible for the new age of growth hackers. The internet has given the world a new kind of product, and it demands a new kind of thinking.
For the first time, because of this redefinition, a product can play a role in its own adoption. Sound crazy? It is. A product like Facebook allows you to share their product with other friends to make your own experience on their platform better. Shampoo can’t do that. A product like Dropbox can give you free cloud storage if you get a friend to sign up with them. Couches don’t do that. If you don’t come to grips with this new classification of products that the internet has produced, you won’t fully grasp growth hacking.
Sean Ellis, that guy that coined the term “growth hacker,” was also the first person in charge of Dropbox’s growth. He understands what is new about internet products. Just look at the growth scheme in this screenshot:
Growth hackers understand the latent potential of software products to spread themselves, and it’s their responsibility to transform this potentiality into a reality.
Despite the importance of product, it would be foolish to restrict your activities to only the product. The same internet that redefined product has also redefined distribution, and not all distribution is within the product. Those with a strong understanding of how people flow online will be able to use that knowledge for the sake of their startup’s growth.
Consider the highway system built in America starting in the 1950s. McDonald’s understood that the interstate roads were a new channel for getting customers, and they took advantage of this. Exits are littered with golden arches to this day. This was an example of off-line growth hacking (if there is such a thing).
The internet is the modern counterpart to this analogy. If you grock the invisible online maps that now direct people, data, and ideas, then you can setup your own golden arches where you know they will be seen. Here are some basic examples to get you thinking in the right direction:
- Instead of highways providing a way to brick and mortar businesses, we have search engines providing a path to digital businesses. Those who master SEO are being seen by everyone who drives by digitally.
- Instead of roads that lead us to local movie theaters, we choose to browse YouTube. Those that truly understand this will be able to get eyeballs on their product in a number of different ways.
- Instead of streets providing a way to our friend’s house, we opt to socialize using Facebook. Those who are aware of this will be able to inject their own agenda into the conversation in implicit and explicit ways.
There are many more examples of the online infrastructure that is creating massive opportunities for product distribution, but the point is that those with an accurate notion of how people move about online will have growth advantages that are hard to imagine. Above are the examples that we all sort of get, but there are hundreds of other examples that take work to uncover, and that’s where the notion of a hacker comes in.
What does the “hacker” in growth hacker mean?
The word hacker has a few different definitions and connotations that inform the meaning of growth hacker.
Hacker is sometimes used to refer to someone who is clever, original, or inventive. They will use whatever is at their disposal to create a solution that might have been overlooked by others. A “life hacker” would be an example of this use of the term. This same attitude is found in growth hacker because they are forced to be ingenious if they are going to achieve growth. Paths to growth are not usually obvious and it takes extreme creativity to find them.
Hacker is sometimes used to refer to a software engineer, and while a growth hacker may or may not be a programmer, they use technology based solutions to achieve many of their goals. Growth hackers will use software, databases, API’s, and related tools to grow a startup. If a growth hacker is also a programmer they can sometimes make progress more efficiently, but this isn’t required. However, a growth hacker must understand technology very deeply to be successful. If a growth hacker isn’t a programmer they will still have to understand programming enough to coordinate others who do write code. Remember, products are now technology based, and mastering the technology will be essential for growth.
Hacker is also used to describe someone who gains unauthorized access to a system. They break into something without permission. A growth hacker will not hack in the illegal sense of the word, but they they will push the boundaries of what is expected or generally advised. A popular idea within computer hacking is “zero-day exploits”, which are security holes which create instant vulnerabilities once they are known. There are zero days between the knowledge of the security hole and the exploitation of the security hole. Likewise, a growth hacker will take advantage of similar kinds of exploits. When a new social platform releases an API a growth hacker might use it to gain users before the API is “fixed” to close the hole they used. Growth hackers are on the lookout for system weaknesses which will allow growth.
What does growth hacking look like in practice?
Up until now we’ve been talking very philosophically about growth hacking. We went through its history, its definition, and what makes it new in the marketplace of ideas. But I know what you’re thinking — give me an example!
In one sense the rest of this guide will be concrete examples, but here is one popular case study that we can use to wrap our head around growth hacking. It’s none other than the poster child of growth hacking, AirBNB. As many of you know, they allow anyone to convert their spare bedrooms into a hotel room that can be rented by perfect strangers. It’s an amazing idea, the execution is incredible, but growth hacking is what possibly put them on the map (pun intended).
They leveraged Craigslist, a platform with millions of users looking for accommodations, to increase their user base substantially. When you are filling out the form to list your bedroom on AirBNB they give you the option to also post the listing to Craigslist, so that it will show up there also, creating inbound links for you and for AirBNB as a platform.
This seems so obvious in retrospect, you may wonder why other companies hadn’t already saturated Craigslist with these kinds of cross postings, making it a noisy channel for customer acquisition. Good question. The answer lies in the fact that Craigslist didn’t have a public API. In layman’s terms, Craigslist didn’t offer an easy way for other companies (like AirBNB) to post to their service. There wasn’t a technological solution that AirBNB could implement easily, and there definitely wasn’t any reference documentation that AirBNB could use to make their listing appear on Craigslist automatically. Instead, they had to reverse engineer how Craiglist’s forms work, and then make their product compatible, without ever having access to the Craigslist codebase. API’s are easy. Reverse engineering is not.
Using this case study, think about how our philosophical meanderings from earlier actually ring true.
- First, AirBNB did something that a traditional marketer would have a hard time envisioning, much less executing. A bachelors in marketing, as it is currently being taught, is not going to give you the tool set, or even the conceptual framework, to arrive at this sort of deep integration with Craigslist, especially sans API.
- Second, AirBNB used their product as the primary means of distributing their product. The integration with Craigslist wasn’t something external to AirBNB’s app. It was a part of it. They didn’t run magazine ads to drive traffic to their product. The product drove traffic to itself.
- Third, AirBNB realized that the distribution mechanism that they needed to hijack was Craigslist. No product exists in a vacuum and the users they needed were already congregating in a different location. So they got their attention.
- Fourth, they were ingenious. They didn’t read about someone else using Craigslist to cross promote something. They thought of it themselves. Then they had the guts to execute on a beautiful solution when there weren’t any guarantees that it would actually work.
- Fifth, their growth mechanism was heavily technology based. The team at AirBNB that pulled off this strategy had to have a lot coding expertise, and a general understanding of how web products are built in order to reverse engineer Craigslist.
- Sixth, they took advantage of holes in an existing marketplace to acquire users. Craigslist didn’t create a public API for a reason. Craig Newmark doesn’t want you doing this on his service. AirBNB pushed the bounds of what is acceptable by not asking for an API, and moving forward without one.
In fact, it looks like Craigslist has “fixed” the vulnerabilities which allowed this integration. Now there is a FAQ answer on AirBNB’s site that says they no longer post to Craigslist. This serves as a great object lesson for growth hackers. Most growth mechanisms have a finite lifespan. It would be unwise for AirBNB to assume that they could post to Craigslist for the next 10 years, as if Craig would allow them to siphon off users little by little every day. But that’s ok. Taking advantage of this temporary opportunity gave them a base of growth that they could use to propel themselves forward.
The future of internet businesses
Growth hacking is an interesting trend that gives us glimpses into the future of internet based companies. There has often been a barrier between the product team, and those responsible for acquiring users for the product. The coders build. The marketers push. It seemed to work for a while that way. Now, those in charge of growth are having to learn what an API is, and those in charge of programming are having to think about the customer experience within the product. Worlds are colliding.
This cross pollination makes sense. If growth really is the lifeblood of an organization, then why wouldn’t growth be woven into every aspect of the organization. Even customer support should be done by people that think about growth because angry customers churn. And designers should design with one eye on growth because beautiful art alone doesn’t always acquire users. The future of internet companies, and the teams that build them, will not look like they did yesterday.
One more note on the future. For now growth hacking is relegated to startups, but eventually, growth hacking will be a part of fortune 500 companies. Startups generally lack resources, and the established relationships, that would allow them to be effective with the tactics of a traditional marketer, so they are somewhat forced to growth hack. However, there is nothing about growth hacking that cannot be applied to larger corporations. If growth hacking can work without resources, imagine what it can accomplish with resources.
The Profile of a Growth Hacker
As this new world of growth hacking comes to prominence, and jobs begin to open up, individuals who are enticed by the possibility will wonder if they have what it takes to be a growth hacker. As with any career, certain kinds of individuals will flourish more than others, but before getting into that, let’s debunk a few myths.
You have to be a programmer to be a growth hacker
Gagan Biyani, the co-founder of Udemy and a growth hacker himself, said, “many of the growth hacking descriptions on the web are unnecessarily restrictive. I don’t believe growth hackers must be formal engineers when many of the most well-regarded growth hackers don’t code regularly.”
It’s understandable why this mistake is sometimes made given the dependence upon code to achieve many growth hacking goals, but it just simply isn’t true that a growth hacker needs to be a programmer. A growth hacker usually needs a programmer on his team, but he doesn’t have to be the programmer. Consider the following example, which is a composite of actual situations:
A growth hacker at a small startup has a team of three people: himself, a front-end developer, and a back-end developer. The team has just been assembled and no growth hacking has taken place. The two developers have never really thought about growth much. They know it’s important, and they are excited about learning, but until now they’ve never been on a growth team of any kind. The first day they are together the growth hacker talks to everyone about event-based analytics and why they matter. The growth hacker then makes a list of every event he wants tracked on the product (signups, referrals, interactions, etc.), and then he shows them KISSmetrics, his tool of choice. He tells the developers where to find the documentation which will help them track events on their product, and then he leaves them to implement the code.
Some weeks pass, and the growth hacker calls everyone together. He uses the data that has been coming in to build a funnel within KISSmetrics. He teaches them what makes a good funnel, and he lays out some goals for each step of the conversion process. The team decides to focus on the unique visitor to signup conversion rate. The growth hacker’s hypothesis is that people are bouncing instead of signing up because the copywriting isn’t clear enough, or powerful enough, so he crafts some new copy, and the front-end developer makes it look beautiful on the site. The conversion rate goes up by 7% for that part of the funnel, in the most recent month over month cohort.
This team, as a whole, is now responsible for growth hacking, but ask yourself this question. Which person, of the three, was the fuel that made growth a real possibility? The engineers alone would not have increased conversions by 7%. Of course, some programmers could do everything in this story, but that’s not the point. This is a thought experiment to show that a growth hacker doesn’t have to be a programmer.
Traditional marketers can’t be growth hackers
Growth hacking has become a sort of religion, in a bad way, and marketers are viewed as the opponent, instead of a very close ally. If anything, a growth hacker is a marketer which has restricted their activities to growth alone. Yes, this focus has created a subculture which looks less and less like marketing as time passes, but their roots are not diametrically opposed.
Consider the story from above about the growth hacker and the two engineers. Traditional marketers usually have copywriting within their skillset. So they have a leg up on someone without any training. In the example above, if a marketer took their copywriting skill, then narrowed their focus to growth alone, and implemented event-based analytics in order to run funnel analysis for different cohorts to track improvements due to copywriting, then they would be growth hacking. The traditional marketer is best poised to become a growth hacker (compared to the general population), should they choose to, if they have a strong analytical and technical mind.
In fact, Sean Ellis, the godfather of growth hacking, called himself the first marketer at Dropbox. Marketer was the title he most associated with when there wasn’t another label for his role. Many of the best growth hackers working today continue to sport the title of Chief Marketing Officer or VP of Marketing. More and more companies are choosing to have a Growth Lead, a VP of Growth, or even a Growth Hacker, but they used to just be called marketers, so let’s not forget our roots.
You Have to be Unethical to be a Growth Hacker
Whenever you narrow your focus to one singular goal (in this case, growth) then you run the risk of making decisions that are not in the best interest of others. Every growth hacker must draw the line somewhere, and like any discipline, it will have bad actors.
In recent months Path has been accused of going too far because of their aggressive practices around obtaining the phone contacts of their users, and the way in which they messaged those contacts. Many think they went too far. I tend to agree. However, I don’t think that AirBNB went too far because they were actually serving people while they served themselves, and Craigslist could change their site at anytime to disallow AirBNB’s actions. Path users can’t undo the spam that went out in their name.
But here’s the real point. Most growth hackers don’t even have to ask the question of what is ethical. They are building harmless product features that increase conversions, and they are getting that product into the public through their knowledge of distribution channels. It’s smart, not unethical. Every growth hacker has to decide if they are going to be a Jedi or a Sith.
Growth Hackers are Extremely Analytical
Ok, we’ve talked enough about what a growth hacker isn’t, so let’s talk about what a growth hacker is. One of the core aspects of any growth hacker, whatever background they come from, is their love of, dependence upon, and understanding of, analytics.
Analytics is the blood that flows through the veins of a growth hacker. Almost everything they do has an element of analytics either in the foreground or the background. Without analytics a growth hacker feels naked. Here are some of the ways that growth hackers use analytics:
Analytics keep growth hackers honest
The world of marketing has been a place of feelings and emotions for quite some time. What was the ROI of the billboard in Times Square? Who knows, but it looks cool, right? Times have changed. Now it doesn’t matter how charismatic you are in a meeting, or how powerful your ideas seem, or how many sheep in upper management signed off on the campaign. The analytics will uncover your awesomeness or your daftness. Period.
Dan McKinley, a principal engineer at Etsy, tells a great story about their infinite scroll fiasco that sums up this point well. After spending five months creating an infinite scroll for Etsy products, they released the new feature. Of course, they celebrated first. High fives. Back smacks. The usual. Then the numbers came in, and people were buying fewer things through search. What? Needless to say, they got rid of the infinite scroll. Here are two of the slides that Dan uses during his talk on this. Hot isn’t enough:
There are a lot of takeaways from this anecdote, but for our purpose we’ll focus on the analytics. If they hadn’t depended on analytics they might not have realized their error. Analytics keep growth hackers honest.
Analytics shift the focus of growth hackers
When you have systems that are tracking your product and activities, the numbers have a way of shifting your focus in unforeseen ways. You might have never dreamed of spending more resources on your referral loop. It might have been a throw away feature that you put in the product just to see what would happen. Then, after you dig into the data, you realize that over 20% of all new signups are coming from this loop, and their lifetime value is higher than your average user. You know that you can make the loop much more efficient, so you change the focus of your team for the next two weeks to focus on this feature. Analytics can help stack rank your to-do list in interesting ways.
Analytics make success repeatable
When you don’t take analytics seriously you can’t efficiently repeat past successes. If all you know is that the company made more money in Q4 than in Q3, then you know nothing. Why was Q4 better? Were there more users signing up for your product, or did you just convert higher numbers of those that did sign up. Was there a particular feature that began to be used because of a recent redesign? Did the AdWords campaign finally began to have a positive ROI because a competitor who bidded up the cost per click stopped running Google ads? If you know what is leading to your success then you can repeat what is working (and stop what isn’t working).
Analytics predict the future for growth hackers
Companies make bets on the future everyday. They guess what the competition will do. They guess what the market will want. They guess at ways to skate to where the puck is going, instead of where it’s been. They guess. Let me be clear, the future will always be a guess to some extent, but inductive reasoning based on analytics allow us to make informed decisions about tomorrow, based on yesterday’s data.
Will the sun rise tomorrow? Technically, there is not a deductive way to know, but inductively we can reason that it will since it always has. When you look at your charts and there is clearly a line moving in a particular direction there is no guarantee that it will stay the course, but if other factors remain the same, it probably will. This isn’t an exact science, but it’s better than guessing.
This is also where correlation and causation become important concepts. If your analytics show that A and B follow a similar course then this information could be used to change the trajectory of those stats. You could run experiments to see if A and B are just corollaries of one another, or if one of them actually causes the other. When a growth hacker uncovers causation, a process heavily aided by analytics, they have a very powerful weapon at their disposal.
Growth Hackers are T-shaped
When it comes to the skills possessed by a growth hacker, they need to be shaped like a capital T. Here’s what I mean. The flat horizontal part at the top of the T represents all the various skills and disciplines that a growth hacker needs to be familiar with. You need to know at least something about many different things. You need to know a little about psychology. You need to know a little about viral loops. You need to know a little about drip email campaigns. You need to know a little about…well, you get the picture. There shouldn’t be anything mentioned in this book that you couldn’t hold a conversation about.
But that’s not enough. You also need to have a few skills that create the vertical line of the T. These are the skills where you dominate. You are the expert in these areas. You go deep. Maybe you know everything about onboarding and 85% of everyone that signs up for your product gets to the MHX (must have experience) which keeps churn down. This can make up for a lot of mistakes. If you can take any piece of the funnel and make it drastically more efficient then you have something to begin building your growth around. You need a few things that you’re awesome at to even have a chance at scaling.
However, here is what separates the professionals from everyone else. Professionals are not happy with a T-shape. They want a V-shape. As they begin to master more and more disciplines they don’t have one or two vertical lines representing deep knowledge, but rather 10 or 20. This creates a V.
Growth hacking seems mysterious, but it really isn’t. A growth hacker is less like an illusionist and more like a marathoner. There’s no smoke and mirrors here, but rather a lot of hard work to master the skills that pertain to growth. If you want to finish a marathon then you have to train for months in advance. More than that, you have to train in the right way. Likewise, if you want to grow a product you need to become a T-shape, then a V-shape, and who knows, maybe eventually a U-shape, and this will take months of training. There is no shortcut to running a marathon or growing a product.
Growth Hackers are Also Right-Brained
Growth hackers spend so much time talking about analytics (truth #1) that it’s easy to forget that they are also heavily right-brained. Analytics are super important, but so is curiosity, creativity, and a general fondness for anecdotal evidence and qualitative facts. If I were forced to select between analytics and anecdotes, I would choose analytics. Luckily, I don’t have to make such a choice, and there is no need for a false dichotomy. Products are machines that usually fly when you have the right mix of whimsy and hard science. Don’t err in either direction.
Curiosity killed the cat, but…
A lack of curiosity will kill your product. Growth hackers have an urge to think new thoughts and try new things. If you want to follow a manual that outlines every procedure for your job then become a middle manager. If you want to follow orders then join the army. If you want to grow a product then get curious. Maybe something has never been tried until now because it’s a dumb idea, but maybe it’s never been tried because no one has ever been curious enough to find out if it would work. Here are some examples of curiosity…
- What would happen if we made our entire product invite only, and not just for the beta period?
- What would happen if we made our users do something every week to keep their account from being deactivated forever?
- What if we let our customer’s pick their own price, including free?
- What if we gave away an upgrade to our product to anyone who pays to have Chinese food delivered to our office during lunch hours? And we tweeted about it every time it happened.
- What if we made our entire homepage an homage to the heroes of our industry on the 1st of every month, and we made it easy to share with friends?
- What if we went through and rewrote all the error message copywriting on our site using famous quotes from cult classic movies?
- What if customer support requests triggered a drip email campaign of hilarious videos from YouTube pertaining to their problem?
Are these ideas stupid? Many of them probably are, but at least I’m brave enough to write them down in a book that will be read by thousands of people. Why are you afraid to write down dumb ideas in a text file that no one will ever see? Curiosity is a function of overcoming fear. Fear of being wrong. Fear of being right. Fear of being different. If you don’t have the guts to think about really bad ideas, you’ll never have the opportunity to execute brilliant ones.
The squishy, ephemeral, fluffy stuff is not your enemy
Logicians and mathematicians crave a binary world. Everything would be yes or no. The data would be clear. The plan would obvious. But alas, we live in a world of grey, where “sort of” and “maybe” are the answers to many questions. The growth hacker must never forget this.
The data may be showing a drop in conversions from the second to the third screen in the checkout flow. You can stare at numbers for hours and still be lost, or you could go to Starbucks and ask a total stranger to checkout using your credit card. I bet you’d learn something. Sure, one person isn’t a large enough sample size to make statistically relevant decisions, but not all problems need mountains of data. You might realize, after watching Jenny from Starbucks trying to check out, that you didn’t give a description of what “CVV” stands for and that she had no idea what to put in that field. How much data do you need to put a helpful message on the site that corrects this? Sometimes one anecdote is enough.
Growth Hackers are Obsessive
Do you really have what it takes to think about growth all-the-time? Sure, it’s fun for a week, but will it be fun in 6 months? Do you have the capability to focus on a narrow goal to the exclusion of everything else, for the foreseeable future? Here is why it’s important to be obsessive:
It’s the 213th tactic that will probably work, not the 7th.
If growth hacking was just a matter of trying five or ten things and then watching the users signup and the money roll in, then there would be no need for a book like this. The truth is that growth hacking only looks simple once you’ve found out the things that work for your product. Until then you have to try hundreds of dead ends.
With enough papercuts you can kill your competition
There is sometimes the assumption that all you need is one breakthrough to win. One big awesome growth hack to own your market. I do think that you can kill your competition, but it usually occurs because of a million micro-lacerations, not one huge one.
Small successes compound over time. If you are able to stay the course and improve your numbers day by day, then you’ll look up after a year and realize that you actually moved the needle in some pretty remarkable ways, but there might not be a breakthrough moment. Bryan Goldberg, writing for Pando Daily, said:
“Whenever I pitch a VC, one of the most common questions I get is this one: “When did Bleacher Report really take off?” The answer is never. As of today, Bleacher Report is one of the 50 largest websites in the United States…But what is even more fantastic is the chart of how we got there. Now, I challenge any reader to pull out a pen and put an “X” over the spot in which Bleacher Report achieved escape velocity. What you may find is that it cannot be done. There was never a moment in which we “took off like a rocket.”
The Growth Hacker Funnel
If you’ve ever put oil in a car then you know what a funnel is. A funnel has a wide opening at the top and as oil runs down it (sticking with our car analogy) the opening becomes smaller and smaller until the oil reaches the engine, which is the ultimate goal. A funnel is a way to guide something which is usually unwieldy and uncooperative, like liquid.
If you are building a product then your task is to guide people towards a particular goal (signup, checkout, etc.). The problem is that people are unpredictable and full of free-will. If you are going to get people to do what you wish, en masse, then you must employ a funnel. When you think about growth hacking the image of this funnel should dominate your understanding:
Defining the three levels of the funnel
The first goal in the funnel is to get visitors. This is the act of getting anyone to visit your website or app for the first time. They are called visitors at this stage because they don’t belong to you yet. They haven’t opted in to anything. They aren’t members, or users, because that would imply that they have some sort of relationship to you, and they don’t. They are just strangers that just happen to be on your site. They are visitors. There are three, and only three, ways to get someone to visit your website or app. You can pull them in, push them in, or use the product to bring them in.
Getting a visitor is like going on a blind date.
After a visitor lands on a site, this is when rookies think they’ve done their job as a growth hacker. Not even close. Now you have to activate them and turn them into members. An activation happens when they have taken an action, large or small, that creates a relationship with you. This might be joining an email list, or creating an account, or even making a purchase. You could even have multiple activations that you track. Now, they are not just visitors, but they are members. They have joined what you are doing in some way.
Activating a member is like being in a relationship with someone.
It’s hard to turn a visitor into a member, but it’s even harder to turn a member into a user. A user is someone who, as the name implies, uses your product regularly. This is someone that you’ve retained. You’ve kept them around for a period of time. If you create retained users then you’ve reached the holy grail of growth hacking.
Retaining a user is like getting married.
What are good conversion rates for this funnel?
One of the great difficulties of utilizing this funnel for your company is knowing what good conversion rates actually are. As you move down the funnel, less and less people stick around. In a given month, you might get 100k visitors, but only 1k members (1% conversion), and only 700 retained users (70% conversion). Are these numbers good? It’s almost impossible to know for a number of reasons:
- Is your traffic from a source that would identify with your product, or are they people that should bounce as soon as they read your first headline. Certain traffic sources will always convert to members at higher rates.
- Does your activation goal include a purchase or are you simply trying to get an email address on file? The more you ask for the lower the conversion rate will be.
- In terms of retention, does your market usually experience high retention rates, or would it be an anomaly to have repeat users? Likewise, is your product a consumer web product that should expect to have incredibly high retention if it’s going to survive?
Given all the variables that go into knowing whether you have good conversion ratios through the funnel, here are some tips to keep in mind:
- Your numbers should always be improving, or you’re doing it wrong. Despite all the unknowns, you should at least be improving month over month relative to your own historical performance.
- Some companies publish their conversion ratios for certain aspects of this funnel. If you compile enough of them then you can begin to benchmark your performance against their metrics. There is a good study of the conversion rates of over 100 SaaS companies at: https://blog.totango.com/wp-content/uploads/2012/11/2012-SaaS-Conversions-Benchmark2.pdf
- The buddy system works well as you think about your funnel. If you can find another growth hacker that has a similar (but non-competing) product, then you can both agree to open up your numbers for the other person. This is one of the best benchmarking tactics for understanding the success or failure of your funnel conversions rates.
- Ratios throughout the funnel are not siloed. You might do something that drives up visitors by 1,000%, but by doing so it drives down retention by .05%. If you make this change and then dwell on the fact that your retention dropped then you’d be missing the point. The retention ratio is going down, but the number or retained users is actually going up. Your goal is to create conversion rates throughout all the stages of the funnel that work together to create the largest overall impact. Don’t miss the forest for the trees.
Let the funnel set your growth hacking priorities
As you consider where to place your energy, the funnel can sometimes make this decision for you. If you are converting 50% of all visitors to members, and 50% of all members to users, but you are only getting 200 new unique visitors a day then you should obviously spend your time getting visitors. In other situations you might want to wait on getting visitors until you are more successful at moving people through other aspects of the funnel.
On a related note, Sean Ellis has popularized the idea of product-market fit, which has a lot of value as you decide your priorities using this funnel. Sean has often said that if at least 40% of your existing users wouldn’t be “very disappointed” if your product disappeared then you don’t yet have product-market fit. This basically means that your product doesn’t solve enough of a pain. It isn’t adequately loved by the users, and the team needs to focus on product more than growth. His overall point is that you shouldn’t try to find new visitors, or optimize the funnel for them, until you have a product that people actually want.
This creates a little bit of a catch 22. If you don’t have any traffic then you won’t have users to poll as you try to find out how disappointed they would be in your absence. However, focusing solely on growth would be a bad move, as you’d be optimizing in vain if your core offering is lacking.
Therefore, here is what I recommend. Use this funnel to get an adequate user base. Then, circle back to them to find out how well your product fits their needs before taking things to the next level. You have to grow some to know if you’re even on the right path to grow more. Just don’t put yourself in a situation where you are expending massive energy in an attempt to growth hack a product that people don’t love. It’s that simple.
It all began with pirates
I need to give Dave McClure some credit. A few years ago he started presenting a slide deck that he called Startup Metrics for Pirates. What did these metrics have to do with pirates? Well, the acronym he used to present his material was A.A.R.R.R. which stood for Acquisition, Activation, Retention, Referral, and Revenue. This framework has been celebrated and for good reason. I would be lying if I said that it didn’t deeply influence my thought processes. Therefore, the framework that I use is similar (get visitors, activate members, retain users). Here are the reasons that I prefer this simplified funnel as opposed to Dave’s funnel:
- It uses the words visitor, member, and user, which actually corresponds to a person’s state at different parts of the funnel.
- Referral (the first R in Dave’s framework) is really just another way of getting traffic. Therefore, I simply the funnel by including product referral mechanisms as a subset of getting visitors, not it’s own category. Also, referral is only one way to use the product to gain new visitors, as we’ll see.
- Revenue (the second R in Dave’s framework) is really just a kind of activation. If you choose your activation step to be a purchase of some kind then it doesn’t need to be another step in the process. This will help us think about activation tactics in reference to revenue more easily.
- It’s simpler, but doesn’t lose any of the organizational power or depth of insight.
Tools and Terminology
It may seem like a section dedicated to the tools and terminology of a growth hacker should be closer to the beginning of this book. Tools are so basic, so fundamental. Terminology is the shared jargon we use to communicate. Well, there is actually a good reason why I’ve saved this section for the end. Tools and terminology can make people feel like they are growth hacking when they have no clue what they are doing.
If you just install an analytics tool, without a philosophy on what growth hacking is, or an understanding of how growth hacking works, then you will just be playing with numbers, randomly changing things in your product, with a false sense of grounding. The metrics will make you think you know what you are doing. The data will look so official. The charts will look so alluring. You’ll be using complicated words to describe what you’re looking at. Don’t do that.
Another reason I saved this section until now is because the tools are always changing, but the process of a growth hacker doesn’t. Whatever tool you use you’ll always be guiding people through your funnel. You’ll always be creating hypotheses, running experiments, and optimizing the results. However you get the data, and however you label the data, you’ll always be engaged in a process that is deeper than the tools and terminology you use. Growth hacking is a mindset and process more than it is a collection of tools.
In fact, growth hacking is not just a collection of tactics either (despite the number of tactics outlined in this book). Growth hacking is the process that arrives at those kinds of tactics. Sure, use the tools, use the terminology, and use the tactics, but understand the fundamentals of growth hacking before jumping to the conclusion. If you allow yourself to think like a growth hacker then you’ll uncover tactics that no one else is aware of, and you’ll be able to survive even when the current round of tips and tricks have run their course.
This is far from an exhaustive list, but following are some of the terms that you will hear often in growth hacking circles. I’m not going to give a confusing, super precise definition. These are going to be defined in plain english. I am willing to sacrifice technicalities for clarity:
Key Performance Indicator (KPI)
Every business has different KPIs, but every business has them. A KPI is a number that helps you get a quick grasp of how things are going within your company. If you are selling software subscriptions then a KPI would be how many new subscriptions you have sold today. Another KPI would be how many people canceled their subscription today. A KPI is not an obscure data point that doesn’t have meaning unless it’s ran through a complicated equation. A KPI is a number that matters for obvious reasons, and by simply looking at you can get a sense of company trends and company health. Here are a few best practices around KPIs:
- Have an automated email send out KPI data every day, week, and month.
- Have a dashboard that has KPI data displayed in such a way that you can see trends based on past performance. It’s helpful to see if a KPI is going up or down in general.
- Allow everyone in your company to access KPI data. This will inform people about which metrics matter to the company, thereby influencing their decisions.
Viral Coefficient (K)
The viral coefficient is a number that tells you how many new people are brought into your product because of your existing users. If every 50 visitors to your product bring in 100 new visitors to your product then your viral coefficient would be 2. Anything above 1 means that you are growing virally. Here some things to know about virallity:
- Going viral is based on an equation. It’s not just a phrase thrown around to describe something that is seemingly everywhere online.
- A viral coefficient over 1 is a great thing, but even if you are below 1 it’s still a benefit to the company. Virallity isn’t always the goal (or even possible). Anything above 0 means that you are amplifying your product distribution to some degree.
- Virality is probably focused on too much. Growth hacking is a large set of skills, and it’s possible to grow a product substantially, and profitably, without worrying about virality.
A cohort is a portion of your users based on when they signed up for your product. Everyone that signed up in January is in the January cohort. Everyone that signed up in February is in the February cohort. It’s important that you use cohorts because otherwise your data won’t be as clear as it could be. If every month your KPIs are improving for new cohorts month over month then things are going in the right direction. If you just look at a single metric, and average it across all users since the beginning of your product, then your data is being skewed by the good and bad of past cohorts, and you are not seeing how things are currently going with your product as clearly as you should.
Segments are like cohorts, but instead of basing the group on signup date, you base the group on other segmenting factors. You might categorize your users into male and female groups, in order to see how they behave differently. You could even break cohorts into segments if this gives you relevant data for your product.
Multivariate testing (or A/B testing) is when you make product changes that are only seen by some of your users. This gives you some people that see the A version of your product and other people that see the B version of your product. Then you can see if version A or B gives you the results you want. A debate within multivariate testing is whether or not multi-armed bandit testing is the best kind of A/B test. Bandit testing is a continuous form of A/B testing that always send people toward the best performing options. In essence, the experiment never ends. I’m not going to get into this debate here, but I wanted you to know that there is a debate.
When running A/B tests it is important to remember that sample size matters. If option A and B are only shown to a few hundred visitors then it doesn’t really matter what the results say. You don’t have enough participants in the test to make the statistics meaningful. Here is an anecdote to illuminate this point. If you run an A/B test for 1 day, and on that day you have 2,000 visitors, and option A gave you the results you wanted 70% of the time, then you’d think this experiment is a success. Later you find out that this was the same day that a new blog linked to your product and the traffic from this blog accounted for 90% of your traffic that day. At best, your A/B test has uncovered something about the audience of this blog, not about the users of your product as a whole.
A nifty web app that let’s you calculate experiment variables which are relevant for a successful test can be found at www.experimentcalculator.com.
Customer Acquisition Cost (CAC)
Customer acquisition cost is the amount of money it takes to get a new customer. If you spend $500 on Google Ads and this gives you 2 new customers then your customer acquisition cost for this channel is $250. It’s important that you know the CAC for each channel because it can very greatly. Also, once you know the CAC per channel then you know how much you can spend on that channel, or if you should spend anything on that channel.
Lifetime Value of Customer (LTV)
The lifetime value of the customer is the expected amount of money you’ll make on someone throughout their entire lifecycle on your product. If people pay you $300 a month for your product, and stay customers for an average of 2 years, then your LTV is $300 x 24 (months) = $7,200.
Segments come in handy when calculating LTV because you might discover that certain segments of your users have a much higher LTV than other users. This will also affect the CAC that you are willing to pay for those specific segments.
If you don’t have a strong grasp of your CAC and LTV then it is very difficult to use any push tactics for getting traffic. It is also hard to forecast your financials, which can have an impact on hiring and other initiatives not related to growth hacking.
Now that we are familiar with some of the growth hacker jargon the question is, how do we get this data. For that we turn to the tools that growth hackers use. The most common analytics tools used by growth hackers fall into a few broad categories:
Google Analytics is by far the most popular analytics platform for general data. It’s free and it’s very powerful. Google Analytics is best for high level overviews of your product, but it is difficult to use for granular event or people based analytics (like the kind of tracking that is becoming popular). If you want to watch geography data, device data, bounce data, and other common metrics, then this is a great tool. However, if you are trying to figure out if people churn less after they watch your demo video (a kind of event/people based analytics) then you are going to have difficulties. It simply wasn’t built to provide this kind of information. Most companies have a number of analytics platforms running simultaneously, so despite the drawbacks of Google Analytics you should still have it installed.
Event/People Based Analytics
There has been a renaissance of sorts within the tools that growth hackers use. Due in large part to the limitation of Google Analytics, a number of new software products have been introduced which allow growth hackers to track the kinds of information that they are interested in.
- Do people who use feature X have a higher LTV?
- Do users in segment Y have higher engagement with feature Z?
- And almost anything else you can dream of.
Notice how we are now tracking events, and people, and this leads to new kinds of insights. If you just install Google Analytics, but you are not tracking specific events within your product, then you are stuck in the dark ages. Sorry, but it’s true.
Two of the most popular tools used by growth hackers to track events and people are KISSmetrics and Mixpanel.
Another trend in analytics are the platforms which focus on certain verticals or niches. Now you can find analytics tools that are primarily for mobile apps, or primarily for lean startups, or primarily for ecommerce. There are too many of these to list, but it would be worthwhile to do a Google search for specific products geared toward the metrics that matter for your industry.
As much as software products are used by growth hackers, many of them also use solutions that are built in-house. Sometimes it’s just easier to roll your own dashboard for specific use cases than it is to get a product to do what you need it to. This will depend on your internal engineering resources also. If you don’t have the engineers then you might not be able to build a custom analytics platform, in which case you’ll have to use off the shelf products. When possible, I recommend installing Google Analytics, some event based product, and building in-house solutions when necessary. It’s always better to have too much data than not enough.
If you made it this far then I have to assume that you are very interested in learning how to growth hack, so let me leave you with some advice. Growth hacking is a skill set that can be learned, and I hope this book has pointed you in the right direction, and helped you begin that learning process. However, there is only one way to really master a skill, and that is through practice. Reading a book is one thing. Growth hacking is another. If you want to improve your abilities then find a project (large or small) and begin applying the mindset of a growth hacker to it. You’ll learn more from your own successes and failures then you will from those of others. Your own success will feel so sweet that you won’t ever forget it. Your own failure sting so much that you won’t repeat it. If you want to be a growth hacker then go growth hack something.
If you do acquire the skills laid out in this book then you are future proofing yourself because growth hacking is not going away. Some people want to label it as a fad, but it is simply too powerful to disappear. The name may change, but not the substance. Growth hacking is a new way of thinking, and those who ignore it will be handicapping themselves for no apparent reason.
Here’s one final anecdote. Atlassian is one of the fastest growing software companies in the world and they also have one of the most sophisticated growth hacking teams that I’ve ever seen. They approach their experiments with the rigor of a scientist, and they use machine learning models to refine their process over time. They have one group focused specifically on new customer acquisitions, and another group focused exclusively on the funnel once they are inside the product. It’s a well oiled machine. But here’s the rub; they are an enterprise B2B software company, selling products that the general public has never heard of. They have boring products that enable communication and collaboration within software environments. They are not spreading because they are sexy. They’re growth hacking their way to the top. If you ignore Atlassian, and the companies that are following their lead, then good luck competing in the future.
For further information on the topic of growth hacking tactics and strategies I strongly suggest that you also reading the following 7 posts below.