Growth Handbook

 Date: March 28, 2022

Growth header image

Measure some shit, try some shit, test some more shit, throw the stuff that doesn't work. It's not that complicated!

Chamath Palihapitiya

I spent almost 3 years on the Facebook Marketplace Growth team. I helped to grow Facebook Buy and Sell Groups, contributed to growth of C2C Shipped transactions by 10x in less than 6 months, and led a team that increased young adults engagement on Marketplace and on Facebook App.

How did we do that? This post is a brain dump of all some things I learned during that journey and beyond.

Find Product Market Fit First

You shouldn’t focus on growth until you find PMF (Product Market Fit). PMF means - people actually find your product useful. If it’s shit, it won’t fly. By growing too early you may attract people who will try it once, and never come back. Even if you improve your product a lot. You need to focus on people with high intent who need your product the most.

How to determine whether you reached PMF? The most popular heuristic is to survey customers about your product asking them a question how would they feel if they could no longer use your product:

  • Very disappointed
  • Somewhat disappointed
  • Not disappointed

If at least 40% percent of them say that they would be “very disappointed” then you have PMF.

Pioneer (an awesome platform supporting founders) has a nice course Find Product Market Fit that guides you through the process. I recommend checking it out if you want to learn more about what PMF is.

Another way to measure PMF is to look at your retention curves - whether people are coming back to your product or not. If you see a “smiling curve” as described in this article then you have PMF.

Retention - smiling curve

Growth strategies

After you determine that you have PMF, it’s time to grow. There are a lot of different ways to grow. This is a comprehensive, but not exhaustive, list of growth strategies:

Access

Allow more people to have access to your product. This could be expanding to more countries or more demographics (older or younger generations). Is your product launched only in the US? Launch it to Canada, and Mexico as well. Then the EU, Asia, etc. Some Asian countries, like India or Indonesia, are huge markets! Is your product tailored for a particular demographic, e.g., millennials? Consider expanding the value prop to gen X or gen Z. Another strategy to expand access is to enable login through 3rd parties like Google or Facebook so people don’t have to create new accounts. You can also provide logout experience if you always required login before.

Category expansion

This is a very broad strategy and it depends on the type of your business. If you are selling t-shirts, start also selling hoodies. If you are selling books for elementary schools, start selling books for middle schools and high schools.

Friction removal

How many clicks it takes to make a purchase on your website? Or any activity that your user wants to perform? Simply reducing the number of actions (clicks, filling out forms, etc.) that is required to achieve something on your website or app would make your users use your product more. There is nothing more frustrating than 10s of clicks to go through multiple different views, with a lot of unnecessary information in order to achieve something. I’ve seen that removing 1 text input from a form can increase the completion rate by 10%! Imagine if you can increase your sales by 10% just by removing one field from a form your users fill out to purchase your product.

In-app upsells

In order to bring awareness of new features, or simply trying to promote something on your website or app, you can add prompts across different surfaces. Your upsell will be more successful if it’s put in the right context, at the right time. E.g., someone is buying a book, maybe they also want to buy a pdf summary for an extra $1? You might be less successful if someone came to your website to use a real estate investment calculator you created, and you are trying to upsell your Vue.js Tutorial. Check out 9 awesome examples of upsells.

Notifications

The best way to increase your product retention is to send notifications to remind your users about your product. There are multiple notification challens: in-app (shown in notification section), push notification, or email. Push notifications are now also available in browsers! Similarly like with in-app upsells, notifications will be most successful if targeting and timing is correct. You should also keep track of notification metrics like views (how many people have seen your notification), CTR (Click Through Rate) and CVR (Conversion Rate). Conversion Rate tells you how many people performed the desired action (like buying your product) as a result of notification. You may have notifications with high CTR that do not result in any conversions. Check out 7 examples of how different companies use notifications.

Facebook or Google ads

Paid ads are easy to make, but not easy to make cost effective! Julian Shapiro’s Growth Handbook has a good intro overview of using ads as acquisition channels. I recommend checking out “Acquisition channels” (especially channel cost), “Making ads” and “Running ads”. Again, be aware that ads are easy to make, but not easy to make profitable and efficient. By doing small tweaks in content, CTA text or art work you can increase efficiency by 2x, 5x or even 10x. This can make it profitable or not. If you spend $10,000 on ads that generate $4,000 in sales that’s not profitable. You need to improve your ads efficiency by 2.5x to make it break even. Be aware of artificial growth! A few years ago Chamath claimed that 40 cents of every $1 dollar goes to Facebook, Google or Amazon. You can find many courses, books and tutorials on runnings paid ads online. Be aware that they expire fast! Most advanced advertisers learn how to maximize them to their advantage every day, and you need to be on top of these tricks to keep your ads profitable. Here is a sample Facebook ads tutorial and Google ads tutorial. I also like this Google ads beginner’s guide. You should also consider which platform is best for your product. You can also use both.

SEO

Search Engine Optimization. SEO is for making your website pop-up in search results as high as possible. It’s probably an even more advanced growth strategy than running ads, as you need to actively invest in it, and effects are not immediate. I may write a separate post or series of posts about SEO in the future. Google Search Central is the bible of SEO. For a more general guide check Moz’s SEO guide.

Growth Hacks

In addition to larger initiatives described above, you can also try simpler, faster to implements growth hacks:

  • Change Content - make it shorter and show DIRECT VALUE for the user; explain what is user getting by engaging with your product or taking particular action.
  • Change CTA (Call To Action) content or color of text or background.
  • Add some unit/text/CTA to some impactful view - if you want people to take some action, upsell it where you get the most traffic.
  • Add insights or remove not needed insights - people like seeing stats about how they are doing; e.g., if they listed something for sale, a notification saying “your product got X views in the last Y days” can increase engagement. Even without any explicit call to action in content, people tend to engage with the product to find out more. Strava is pretty good about it. Sometimes they send emails, like your monthly summary: “you ran X miles last month”. I found myself going to check more details about my performance in the app very often. It’s working!
  • Improve performance (removing things from page/view helps with perf, adding things hurts perf) - better performance can lead to growth as users can be more productive, and less frustrated when using your app or website.
  • Remove some not needed unit/text/CTA - an often overlooked growth hack is to just remove some of the least impactful or not relevant anymore things from your product. It’s very easy to end up with feature bloat, but product teams rarely audit them and deprecate/remove
  • Prefil form with defaults (e.g., do not ask the user to select a subscription plan, preselect most popular for them) - reducing friction in forms can drastically increase completion rate, which can lead to significant increase in sales.
  • Change number of items shown - if you are displaying, e.g., newest 4 items, consider displaying 6 or just 2 or 3. If you display less, you may score performance gains. Displaying more may lead to more impact as people have more opportunities for engagement.

You can find a lot of different growth hacks in Aladdin Happy’s books TOP 101 Growth Hacks and TOP 101 Growth Hacks 2. Both books contain 202 different ideas. This can help you with brainstorming your own ideas.

What to work on? Use ICE framework

Coming up with ideas never was the biggest problem for my teams. We always had more ideas than what we could execute. Our brainstorm sessions (powered by mural) always produced 100s of ideas. The dillema was what to work on.

The simple framework that can help you prioritize your ideas is ICE Framework. ICE stands for Impact, Confidence, Ease of implementation. You score ideas on a scale of 1-10 in every category, and then rank them by summing it up. I’ve seen a lot of engineers throughout my career focusing too much on impact, while totally neglecting ease of implementation.

Examples:

  • Redesign landing page (Impact: 7, Confidence: 5, Ease of implementation: 3, ICE score: 15).
  • Create a new notification (Impact: 3, Confidence: 6, Ease of Implementation: 7, ICE score: 16).

Do this for all potential projects, narrow your list down to top 10, and start working on the first one. You can keep other ideas in the backlog, but don’t focus on them too much to not get too distracted. You should always be focused on 1 idea, keep 10 in mind, and the rest keep in the backlog.

How to estimate impact and confidence? Opportunity sizing

There is no straightforward formula for estimating ease of implementation. It depends on the engineers, their familiarity with the product, etc. However, you can estimate impact and confidence with opportunity sizing.

For example: you want to upsell a new product you are selling by creating a Facebook ad. You can estimate the number of people who will see the ad. Based on that you can estimate the number of people who will click on it, and the number of people who will make a purchase after visiting the product page.

Let’s say your ad will be seen by 100,000 people. Based on your past experiences, targeting quality or by just guesstimating based on some data found online about funnel performance, you can estimate your ad’s CTR. Let’s say it would be 2%. This gives you 2000 people visiting your landing page. How many people will buy the product? Ideally you can estimate that based on the % of people who buy a product after visiting your landing page. If you’ve seen that 1% of people who land on the product page buy it (that would be pretty good!) then your opportunity size is 1% of 2000 = 20 products.

You can do similar estimation for notifications or other growth channels.

Note: be aware of cannibalization. If you reach the same people with 2 growth channels they won’t produce a sum of expected results. I.e., if someone sees your ad twice you probably won’t double the number of visits. Maybe you will increase visits by 20-30%.

For more about opportunity sizing checkout this, this or that.

Lessons Learned

Doing trumps Planning

Trying things always produce better results than endless planning, and analyzing. Over years I observed that growth is not proportional to how much effort you put into single change, but rather to the number of different things you tried. It doesn’t mean to do some shitty changes and hope it will magically bring users. Rather, it emphasizes the importance of execution and trying different things. No matter how crazy and unlikely to bring growth they are!

MVE - Minimal Viable Experiment

When you are not sure if something is going to work, instead of building full-fledged features try to run a directional experiment that requires minimal amount of work. If the experiment shows some positive results, double down and build a fully fledged feature.

Make your users winners

In the book Badass: Making Users Awesome, Kathy Sierra explains the importance of making your users feel good while using your product. You need to make users feel “successful” when they are using your product. They need to feel like they are winning something while using your product. If there is no win, there is no point of using your product.

Iterate, iterate, iterate

If something didn’t work at first, but you have a strong hypothesis why it should’ve (e.g., data or just intuition), you should iterate. I’ve seen many times people giving up too early. I have also seen when something didn’t work, because there was no trivial bug affecting a large portion of the users. I rarely see something taking off to the moon on the first try. Even if it does, there is probably still an opportunity to iterate and get even greater impact from it.

Do not run too many experiments at once

In an ideal scenario you just run 1 A/B test for your product. As you want to move fast you may need to run multiple experiments in parallel. In that case, you will get the best results if you isolate them. I.e., you experiment with one thing on the minimum population needed to show statistically significant results, e.g. 20% (10% get a test, 10% control treatment), another thing on different 20% of users, etc. What if you won’t get meaningful results with a running test on 20% users? You may rerun it later with a larger population when you won’t have other experiments to run. Running a bunch of 50/50 A/B tests may result in randomizing user experience and you won’t be able to tell which of the changes is causing (or not) desired results.

Use back tests and holdouts

When you try new things, you may run into novelty effects. I.e., some new feature is causing a temporary spike in engagement, but after some time it goes back to neutral or even negative. Running a test for 2 weeks, and making a decision based on data from only week two should be sufficient to remove the novelty bias. But not always. Feature performance can degrade after a month or two. To monitor long term effects of a feature you should consider using back tests. This means launching new experiences for the majority of the population, and keeping a small number of users who still get the old experience. To monitor the compound effect of changes made in some period of time, consider also using holdouts. Keeping a small number of users who won’t get any of the new features that you introduced in a given period of time. This can help you assess the total impact of changes you made over time. Usually there is a cannibalization effect and the sum of all backtests won’t equal the holdout impact. Keep in mind that using backtest and holdouts require a pretty large number of users to get statistically significant results. Thus, consider these techniques only when your product is more mature, and you have a lot of experiments running at the same time.

What didn’t work in the past may work now

Before joining Facebook Marketplace Growth team I often heard: “We tried this in the past, it didn’t work. Don’t waste your time”. In my early days at Growth team, when I was trying to find different ideas, one engineer advised me to look at past experiments that didn’t work. “If it didn’t work in the past, it may work now”. Timing is one of the key elements of building a successful product. If you are too early with an idea you won’t get anywhere. Another factor in addition to timing is execution. There were multiple attempts at Electric cars before Tesla, and there were a lot of different social networks before Facebook.

North Star Metric

North Star Metric is a metric that the best measures value of your product. It could be Daily Active Users, number of transactions per day or number of posts users create on your service.

When you grow your product, you should have North Star Metric and focus all your efforts on growing it. Would some change increase your North Star Metric? Do it. Otherwise - abandon or postpone. This helps tremendously with correct focus and proritization.

For more about North Star metric check out this post by Sean Ellis.

Recommended Growth Resources

If you want to learn more about growth I recommend below resources. Get familiar with them and try out some growth ideas in your product. Remember that more things you try, more likely you will be successful.

Books

Talks

Articles

 Tags:  growth

Previous
⏪ Moving from WordPress to Jekyll

Next
Skeptics Guide to Universe - Concepts Summary ⏩