Whether you’re aware of it or not, you’ve just added another assumption into the model. By including predicted customer numbers, you now have a value for average amount of revenue per customer. But where are these customers coming from? What channels will they arrive from, and how much will each of them cost? There’s not […]
Whether you’re aware of it or not, you’ve just added another assumption into the model.
By including predicted customer numbers, you now have a value for average amount of revenue per customer.
But where are these customers coming from? What channels will they arrive from, and how much will each of them cost? There’s not enough detail here yet.
Above the customer numbers we add in each of the marketing channels we expect to use to acquire customers, making sure to include any extra costs in our budget. This is what we came up with:
Right now, this model is feeling pretty intellectually satisfying. We’ve got lots of testable assumptions about each method, including some implicit assumptions about cost per acquisition for each channel.
So what’s missing? Conversion ratios.
Your customer numbers are converted customers, not leads. Unless you are converting 100% of your new leads into customers (big shout out to all the crack dealers out there), then by definition you have generated more leads than you have customers.
So your model has you driving 50 customers into your application each month via search engine optimisation (SEO)? Good for you.
How many leads did you generate in order to actually convert 50 people into customers? Start off by assuming you will be converting at around 1%.
Anything higher than this in your pre-launch model and you are setting yourself up for failure.
Based on this, to drive 50 customers from our SEO efforts we actually have to generate 4,950 visits to our website.
Did you know that this assumption was in your model, doing its best to stay hidden and out of sight?
The theory of conversion rates should be applied to every marketing channel you have in the spreadsheet, until you come up with the total number of visits you need to drive into your website.
If you do this properly, my guess is the first cut will have you approximating Amazon.com-style traffic just to break even.
But that’s okay, because now you have one of the most important assumptions in your entire model to work from, namely, the total number of leads you have to drive into your website in order for your model to work.
In your model there are any number of assumptions that surround each metric, and you need to do your best to extract what they are and write them down. You need to know:
Cost per acquisition
Total number of leads generated by channel
Conversion rate
Expected revenue per customer
Retention rate of existing customers
The beauty of these is that they are all now testable, which should be the aim of any marketing in the first 100 days after launch.
Throw some money at Google Adwords and see how your assumptions stack up with some real world data to measure against. You might be pleasantly surprised!
Scott Handsaker is a co-founder of Eventarc, an investment backed start-up focused on online event registration and ticketing.
