Being in the trenches for countless startups, I’ve been lucky enough to see why many growth marketing engines don’t work correctly. I say I’m lucky because the issues I’ve seen have taught me an immense amount about what makes a well-oiled, polished growth marketing engine fire on all cylinders. My experience at Postmates taught me more through mistakes than triumphs. I learned how to correctly scale a growth engine while marching us towards an exit.
What’s most fascinating is that there’s a common thread of mistakes that string all of these startups together. Some of the mistakes that happen frequently are performance metrics not being correctly measured, product and growth teams working in silos, having a slow testing velocity, and not thinking about the entire funnel. This is not to say that there aren’t unique problems at each startup, only that there are a few ubiquitous ones. Instead of additional startups having to face these same mistakes, my hope is that this column serves as cautionary advice.
When measuring the success of a campaign, whether on Facebook for paid acquisition or with a retention series on lifecycle, it’s vital to have the correct metrics to take action on. Having the correct metrics tied to campaigns is a foundational pillar in any growth marketing stack. What if performance metrics are inaccurate? And if they are inaccurate, what could be the cause of it? I’ve listed the top three reasons for not having correct metrics below:
One of the most common reasons for not having the correct metrics comes from reliance on self attributing networks (i.e. Facebook and Google) to measure conversions. These platforms will often over-attribute the number of conversions they’re driving, especially when running many other channels.
Take the example above where Sam clicks on a Facebook ad today (Day 1), and a Google ad two days later (Day 3). Sam converts after clicking on the Google ad on Day 3. The conversion should go to Google, right? In this instance, with the industry standard 7-Day Click/1-Day View attribution window, both Facebook and Google would count the conversion because it occurred within that time frame. This is a double count. If there is heavy reliance on the data from ad dashboards, there is a high probability of double counting.
The second most common reason for incorrect metrics is attribution loss. To complicate things further, there are a plethora of reasons why there may be attribution loss. I’ll list some here:
Of these three plaguing attribution loss possibilities, tech stack issues can luckily be debugged and refined. Without getting into the details of a proper marketing tech stack, I’ll just say that this is an area that should be heavily invested in as growth efforts are ramped up.
Privacy is an ever-changing issue nowadays that can’t be ignored. This one unfortunately can’t be completely resolved, but instead, solutions have to be put in motion to model out conversions that are going unattributed. While at Postmates, I experienced a unique situation where we had to turn off marketing spend for a few days. During this time, we saw organic conversions plummet. This is why having a form of organic incrementality scalar is crucial for the inevitable conversion leakage to organic.
If video or display ads are running, be conscious of the user behavior being drastically different for these placements. A majority of the conversion volume won’t be tracked outside of the ad dashboards, with a large chunk being attributed to organic. This is because users don’t click YouTube or display ads as much. Instead, users search for the product or service being advertised after seeing one of these placements.
And lastly, to avoid any mixups with metrics, make sure to have the process and methodology delineated so that it stays consistent and accurate. It’s common to look at metrics in either an uncohorted (actuals) or cohorted way. With uncohorted, you’re looking at all conversions that happen regardless of when the initial touchpoint happened. On the other hand, with cohorted metrics, you’re looking at how many conversions came from the time of initial touchpoint and spend. I may write a column on the differences of these methodologies, as they’re very different ways to measure performance.
As time has gone by, product and growth have become increasingly intertwined because the roles overlap in a variety of ways. Whether it’s iterating on landing page experiences, implementing new growth stack tools, or tweaking the user funnel — these two roles must work together in order to successfully scale. There’s no way around it.
We’re even starting to see growth product manager roles increase because teams are noticing the importance of these two roles working together. At a startup, this may be the CTO or engineer who’s also wearing the product “hat” and who will need to closely interact with the growth manager. By having these two roles work closely together, the amount of testing and adjustment that can happen, in order to continue moving the needle, is greatly increased.
The days of flipping the switch for paid acquisition, lifecycle, social media, content, and letting it automatically run are far from here. While we’re seeing more automation across platforms, the need to continue testing is paramount.
It’s simple. Test more and the results will come to fruition sooner. While the concept is simple, a proper testing framework needs to be put in place, one that defines the amount and type of weekly tests that are being deployed. A sample weekly test plan can look something like this:
Create a testing framework and most importantly, stick to it. The results will follow.
By far my favorite growth topic, which I don’t think gets enough attention, is none other than incrementality. This is the concept of incrementality in a nutshell: if a growth lever is turned off, how many conversions are lost? This is so important because it places a numerical value on growth efforts. When starting and only running one channel or medium (i.e. only Facebook), it’s not as imperative. As soon as growth mediums or channels get added into the mix, the need to see whether Facebook, Google or affiliates are driving incremental conversions becomes important. A common channel that drives low incrementality is Google branded search because those users are already searching the brand and would have most likely clicked on the organic link regardless. I’ve written more extensively about incrementality in a separate column so I won’t dive into the details.
The one-step syndrome is growth marketing’s worst enemy (yes, I just made that up, but that term explains it best). I’ve seen many startups focus solely on top-of-funnel traffic with no regard for the rest of the funnel which includes activation, retention, and referrals. To make every ad dollar worth it at the top of the funnel, your leads must be nurtured to activate initially, stay with you for the longer term, and become excited enough to refer others.
It may sound daunting to set everything up as a startup, but there are many hacky, quick ways to get all of this launched. At the very beginning, go very broad with your strategy and don’t worry too much about the personalization until later on. Use tools such as Zapier that make automations and connecting tools very easy. There are even ways to set up automatic emails for new leads. Leverage services such as Apptentive (mobile) or Trustpilot (web) to increase your reviews and fix issues early on, so that you retain as many users as possible.
In the early days of a startup, every mistake that’s avoided can dramatically accelerate the time to achieving success. These are just a few common growth marketing mistakes I’ve seen; no doubt there are many others. This column will hopefully start to steer you in the right direction towards a polished, lead-generating growth marketing engine.
How does a founder implement a growth framework to scale to their first million dollars in revenue?