Optimizing your growth funnel
You manage to acquire millions of users. Amazing.
999,999 of them don’t make it through the funnel or churn. Now that’s not so amazing.
That’s an extreme example but it shows why optimizing your growth funnel is crucial during the early stages. What does a general growth funnel look like? Although this will vary between startups, at the core, it consists of three major pillars: acquisition, activation, and retention. I won’t go through each of these pillars in depth (there are many columns on this); however, I will be going through some key optimization concepts.
It’s a dance when optimizing a growth funnel, but let’s dive into how to make it the most methodical and strategic dance.
Acquiring the right users
As a startup matures and robust user-level data starts to flow in, understanding which acquisition sources yield the highest propensity for users should be a priority. A bonus would be to measure the most incremental sources as well.
There should be a defined cadence between cross-functional teams (i.e., product, growth, and data) to continuously question the best sources of traffic.
While leading fleet growth at Postmates, I quickly learned to become numb to remarkable upper-funnel metrics. Below is an example of how we prioritized budget allocation with down-funnel metrics:
Indeed seems to be the worst acquisition channel (CPL) but ends up being the best over time (ROAS). These are not actual results and are purely for demonstration purposes. Image courtesy of Jonathan Martinez.
Even though Indeed has the worst CPL of the channel grouping, it nets the highest revenue-producing fleet drivers who are the most active in Y1.
As traffic sources begin to increase and the spend scale rises, testing for incrementality is the next level up from solely measuring ROAS or LTV. In other words, understanding how many users would have been acquired had a specific traffic source been turned off. How incremental is Facebook or a particular lifecycle marketing campaign?
If there’s one takeaway, it’s to double down on the highest LTV and incremental segments as rapidly as possible.
Bucketing users is key
While there are three major pillars in the growth funnel, hundreds of segments are potentially sitting in each bucket. In the acquisition pillar, users may be bucketed by the channels they came through, while in activation, users may be bucketed by the number of days they’ve been waiting to get started.
It’s imperative to slice (or cohort) users into their respective buckets because it opens the opportunity for unique targeting and messaging.
Each bucket can have hundreds of user segment slices. Image courtesy of Jonathan Martinez.
Below are just a few examples to get your mind thinking about user buckets for each pillar of the growth funnel:
Acquisition
Source
Campaign
Ad Creative
Activation
Funnel Step
Time Lapsed
Actions Taken
Retention
User Behavior
Last Sign-In
Pushing users through the funnel
After slicing users into their respective buckets, it’s time to start thinking about how to message each user. At the foundational level, varied messaging by user slice is the largest lever for moving users through a funnel. To put things into perspective, would you email a user who just came into the funnel the same as someone who has lapsed for 90 days? Probably not.
A few tactics for moving users through the funnel:
Persistency with messaging
Multiple forms or engagement
Alleviating common concerns
Continuous AB testing
Of all the tactics above, nothing works better than continuous and well-crafted messaging to drive desired actions from your users. It often takes B2B sales teams multiple attempts at calling, texting, and follow-ups until a deal closes. The same holds true for B2C startups – i.e., create email drip campaigns and lean on the side of over-communication with your users.
Whether devising a strategy for your email campaigns or retargeting on Meta, always test various styles of messaging. Below is a great framework that I’ve used to define the messaging:
What step in the funnel is the user in?
What concerns or blockers does this user have?
A perfect example of this is the fleet funnel we had at Postmates, which would require people to submit their driver’s license and SSN. We had dedicated email drip campaigns for the people who dropped off at the SSN step to alleviate their concerns of entering sensitive information.
Always low-hanging fruit
There will always be low-hanging fruit, whether it’s changing a subject line or simply adding emojis to retargeting ad copy. It starts with the basics of acquiring the right users, bucketing them the right way, and pushing every step of the funnel. As test throughput increases, the level of detail on metrics should increase to continue moving the needle. Out of all the metrics, my favorite at Coinbase was the incremental lift for each campaign as it helped us accurately understand how many users acted because of a specific email.
Every successful startup that I have been at or advised has implemented B.I. (business intelligence) tools such as Amplitude or Mixpanel to find these low-hanging fruit or even the bigger swings. It empowers growth and product teams to create dashboards filled with cohorts to pinpoint and optimize funnel drop-off points.
With a possible recession on the horizon, it further highlights how an inefficient growth funnel can crush a business. You must optimize as if you need to squeeze every single user through the funnel because of a short runway, then you’ll be much further along in the competition.