How to run efficient growth experiments

The main reason startups raise seed funding is to figure out how to grow. Ideally, you try 10-20 different things to grow over 6-9 months with seed investors’ money. Usually you find that 2-4 things drive the majority of your growth. Then at Series A, you ask investors for much more money to supercharge those 2-4 things to grow exponentially.

There is no way to know which 2-4 things from the list of 10-20 things will work - you have to try them all…quickly, cheaply, and with minimal reputational risk. At the same time you are also trying to find product-market fit 🤯 …but we will cover that in another post.

Doing this properly is not easy. I’ve come up with a quantitative framework to help guide you on how to run efficient growth experiments:

In the spreadsheet above, I’ve listed 11 sample growth experiments. Each experiment has three elements scored 1 to 5:

• Time (1 = 15 minutes; 5 = days)

• Money (1 = negligible; 5 = significant)

• Reputation (1 = minimal risk; 5 = risky)

The “Pain Score” is calculated by multiplying all three: Time x Money x Reputation. 

Examples of great growth experiments

What makes a growth experiment truly great is the ability to execute it swiftly and affordably while preserving brand integrity and reputation. Bonus points if you can be creative and stand out. Some experiments I’ve personally run include:

• Cold outreach to B2B prospects via email + Linkedin (high reputational risk but quite cheap to get leads)

• Placing coupons on windshields to advertise our startup (great results but be careful with local ordinances on doing this)

• Posting on niche platforms (i.e. Reddit threads, Hacker News, etc) to draw traffic to our website (build karma first with your handle)

Some of the most famous examples include:

• PayPal giving new users $10 for signing up and $10 for every friend they referred

  • PayPal spent $60-$70M on this program resulting in 5-6M new transactors (these are very good numbers!)

• Dropbox giving an additional 125MB free after you connect your social media accounts

  • Dropbox’s “North Star” variable was MBs used per user on the platform - they went after stickiness, very important in a crowded space (Box, Google Drive) 

Crucially, we need to pick the variable that we are trying to optimize. This is usually the startup’s North Star variable. In the example on the spreadsheet, I picked GMV (Gross Merchandise Value, how much commerce the platform transacts) which is the North Star of most marketplace startups. I’ve taken an educated guess at the GMV uplift from each experiment - you should update these numbers as you run the experiments.

Picking the North Star or OMTM (One Metric that Matters) is highly dependent on two things: 

• What type of startup you are (B2B, B2C, Marketplace)

• How early your startup is (pre-product, MVP, generating early revenue, scaling)

Typical OMTMs are:

• Revenue

• GMV

• Number of meetings booked

• Demo rate conversion

• Number of app downloads

• DAU (Daily Active Users)

Going back to the example spreadsheet, we can clearly see that Experiment #4 (“Sending double-sided referral codes to your friends to share”) is the leading contender based on a low pain score and high GMV uplift hypothesis.

At this stage, I recommend choosing the top 2-3 Experiments and doubling down on resources (time, money, reputation) for ~10 days to see if they can be reliably counted on as growth accelerators.

Remember, the experiments never end. What works now will likely not work 6 months from now. The key is to keep trying and iterating throughout the early-stage startup process until you have a reliable set of 2–4 growth levers you can pull. Then, it’s time to blitzscale 🚀.

Recently, I’ve worked with several startups to map out their Framework for Growth Experimentation - leading to clearly defined growth experiments to execute, assess, and iterate. If you’d like to book a time to talk through your growth challenges and get an actionable plan, I’m available here. 

-Suhas