Growth hackers should (be willing to) advertise (sometimes)
Like every startup, we spend a fair chunk of our time ‘growth hacking’, which is just a cool way to say ‘trying to get more customers’. And (I suspect) like every startup, we’re skeptical of the worth of paid startup advertising. But we think it would be really silly to say “we don’t advertise”. I think paid advertising of some kind has to always be on the table as an option when you’re growth hacking, provided that these two assumptions hold:
- You want people to know about your business (i.e. you’re not in “secret beta” or selling drugs)
- You are capable of writing ad copy that makes people at a minimum feel slightly more positive toward your company than they did before they saw the ad
Those assumptions, of course, should hold for pretty much every startup. If they do for yours, here’s why advertising should be on the table: there will always be a price at which advertising will be worth buying. Anyone who tells you that you should never advertise is either being sloppy with their words (they mean it’s unlikely it’ll be good value) or isn’t an analytical thinker.
Understanding the value offered by an advertising campaign
You can break down what a particular ad offers into three parts: the price of the ad, the behaviour it drives, at the value of that behaviour to your business. If you understand the conversion pipeline of your business, you can figure out whether the price you’re being offered is good value for your business, based on the behaviour the ad drives.
Imagine there is an advertising opportunity available that costs $1,000. It will be seen by 100,000 people, and will generate 1,000 visits to your product website and 100 signups to your product. You know that you convert 25% of your signups into paying customers, who have a lifetime value of $40 each. On that basis, the ad is a marginal call – it’s giving you $1,000 in revenue (25 paying customers at $40 each) for $1,000 of cost. For most startups, you’d take that offer though, because extra users means extra growth (helps your fundraising, momentum, SEO…) and because for the people that don’t become paying customers, at least they saw your ad / website / product.
We are able to discuss that scenario because we’ve plugged in a bunch of numbers about the behaviour the ad will cause. You won’t ever know that for sure in advance of running an ad. This is another situation where running a startup rewards people who make good intuitive calls.
What would you have to believe analysis
The way around the need to predict into the future with limited info is ‘what would you have to believe’ analysis. I came across this a lot in my former work as a management consultant, but it’s an approach you’ll see in a lot of business and finance settings.
Be clear about what you do know
The first way you can get around this uncertainty is by getting clear on what you do know. Most relevantly, you should have a reasonable idea about your internal conversion pipeline (e.g. how many people who sign up become paying customers), customer lifetime value and churn (see this great a16z blog post for a primer on these concepts).
Here is where you have to be willing to push advertising sales people for some transparency. We’ve found it’s very difficult to get people to tell you, for instance, the click through rates for the eDMs or banner ads. Ideally, advertisers would risk share with you (e.g. by being paid for successful signups rather than just impressions of your ad) but at a minimum they must provide the data you need to assess whether they’re offering a price that works for your business.
Surface and challenge ALL of the assumptions you’re making
The next thing you have to do is make sure you’re surfacing and challenging all of the assumptions you’re making in your analysis. Let me give you a clear example of the sort of thing you might otherwise miss: let’s say from the above you know your conversion rate from signups to paid users, and you know your customer LTV. If you use those figures to work out whether an ad is worth it, you’re making the assumption that users who come in through this ad will behave in the same way that your existing users have.
This assumption might be fine to make, but you need to call it out so you can face into it clearly. This is especially the case if you’re advertising to a totally new audience or user group who you might have reason to believe will behave differently.
Play around with what you’d have to believe until you’re comfortable making a call
Once you know what you know, and know what you’re taking a stab at, you need to just play around with you’re assumptions until you’ve got a set of assumptions and an answer you think make sense. It’s important to recognise this is an art, not a science. If you take a mathematical approach and accept whatever answer comes from your assumptions, you’re putting computational value on a process that is intended as a thinking aid. Similarly, if you are just going to change the assumptions until you get the answer you had in mind, you may as well save yourself the time and just run with the answer (but that would be a dumb idea).
Use Qwilr’s “Should I advertise” tool to help you decide
We have gone through exactly the process I have described in this post a couple of times. As a result we decided to build a tool to help us assess advertising campaigns. It was originally a spreadsheet, until Dylan built the amazing tool you can find here. We hope you find it useful!!