- Startups are highly dynamic
- Sub-optimal decision making can be corrected quickly
- Your marketing can't stop
- Your marketing will benefit from diversification
Status quo? What status quo?
Startups are highly dynamic environments so many aspects of the business are rapidly changeable. Tech startups can now make changes to their product in real time, changing pricing structures manually or programmatically, turn advertising off here, on there and so forth. Never before has the status quo been so non-existent.
Couple all of this combined with low volumes of customer data and you have a potential cacophony of information from which to make decisions.
Decision making is just your best guess
Low volumes of customer data terrify me. When you have small sample sizes, outliers carry disproportionate amounts of weight. You won’t even know which data points are the outliers yet, exacerbating the problem. Companies have the ability to make terrible decisions based on outliers.
One question that regularly gets raised is: “how do I know when I have enough data to make this decision?”. It’s a tricky one to answer. The statistician in me wants to delay the response until I’ve got a significance level of 95%. However, startups rarely can wait this long.
It is time to get comfortable with the fact that the decision you’re going to make is really just your best guess. There is a subtle nuance between allowing data to guide you (relying on 95%+ significance), and making decisions that can be justified using the data available. The first one is easy. It is achievable by anyone with the required sample size and a deft hand at Excel. The second one less so. Justifying your decision making requires you to ask the right question of your limited data, get a result, perhaps test the significance (although a lot of the time this might freak you out). Accept the potential for it to be wrong. Regardless of outcome, use the result to further inform your inherent knowledge about the question to make a decision one way or another. When you have low data you need to gather all the information about a scenario possible and use your judgement to make a call.
This is a tough gig by the way - as a startup marketer you’re being asked to make decision based off limited data in a highly dynamic environment. Lots of the time you’re going to make a ‘wrong’ decision. Paradoxically this is where the dynamic nature of startups will help you as it will allow you to correct your course and avoid that Titanic sized iceberg before it is too late.
A section about metaphorical escalators (bear with me...)
One mistake I often see is a company resting on its marketing laurels. It’s a moment when the company has typically found a marketing channel that has sufficient scope to grow the business and the CPA is in line with expectations. At this point there is sometimes a process of optimisation of the channel in question, and in other cases the channel is just left to itself. This is bad.
Visualise this scenario by thinking of two escalators side-by-side going at slightly different speeds. You’re on the faster one and your friend is ahead of you on the slower one. You get on the escalator and immediately start making ground on your friend. You represent your marketing efforts and your friend represents a marketing channel (i.e. Facebook ads, Adwords etc). Importantly, both are moving independently of each other.
When you and your friend align next to each other on the escalator you have a channel that is working efficiently. Now, knowing that you’re moving at different speeds you are aware that you need to optimise this channel. You can do this by taking a step backwards every so often to maintain your alignment next to each other. You can keep doing this for a while, and until you reach the end of the escalator otherwise known as your marketing channel efficiency inflection point (more here), that’s a great scalable channel.
Now what happens if you speed up the escalators? When you do this you’re effectively pushing the channel harder, driving more from it at an accelerated rate. All seems good if you can optimise yourself at the same rate (although there will be a point where you can’t and fall over backwards - don’t try this at home). However, the end of the escalator approaches faster, meaning you have less time in this channel until you reach your maximum scale whilst maintaining efficiency.
If you only have one channel, this looks fairly simple. However, in reality you’re going to need to spool up some more escalators in case one escalator breaks down. Bringing more escalators online is akin to performing marketing channel diversification. Similar to financial portfolio diversification this aims to act as a bit of an insurance policy if one channel decides to stop working effectively. Another way of saying this is “don’t put all your eggs in one basket”.
escalator maintenance contiguous marketing
The steps below should ensure that your marketing never reaches the end of the escalator, and that if your escalator malfunctions you’ve got some backups in play.
These first two steps, A & B, are crucial to perform before you go to market.
Find pockets of high audience density
The next 6 steps should cycle, to ensure you’re adding to your list of channels before any stop working.
Establish a first channel
Begin optimising that channel (to prove it can indeed be optimised)
Channel test - using the rapid-fire technique
Launch new channel
Repeat 1-6 (forever, and ever, and ever-ever...)