Earlier this week, I wrote an email in which I explained the reasons why I was passing on a deal that my team and I had tracked for more than three months.
The eventual reason for letting go of this deal finally dawned on me when I re-did the calculations for the cost it took for this venture to acquire a new customer (or a unit of “new” sales). When I completed this exercise, I could finally appreciate the vast disconnect in the way the founder and I saw the same traction numbers and the valuation for the company.
First, this is how I recalculated the cost of new customer acquisition, starting with a net gross sales number which was done by:
The NGSn number must be positive, and only when if the Gross Sales / NGSn > 2x it piques my interest.
Next is how I calculated Net Sales (new) or NSn:
The NSn number is usually and understandably, negative – profoundly negative when a start-up is experimenting with different marketing strategies early in their development. However, NSn must turn positive before raising your pre-Series A round as it is a clear indicator of achieving product-market fit.
Those preparing for Series A rounds should get to NSn / NGSn > 0.5x as a clear indicator that each rupee invested in marketing delivers an ROI of 2x or more.
Unfortunately for the founder in my example, the NSn number was deeply negative, i.e. in the -0.5x range. I concluded that this start-up was raking in less than the amount of money and effort invested in marketing, i.e. product-market is not yet achieved. The fact that the NSn / NGSn ratio was touching almost -1x in their best sales month made it difficult for me to assign any positive value their traction.
*I say new sales or new customers because usually any returning customers do not (and should not) cost the company marketing rupees. In the case that returning customers cost the company marketing money, then the budget for that should be kept separate. Remove these returning customer costs from each line item from the formula, to ensure that the ratios are accurate with correct data.