Tren Griffin+ Your Authors @trengriffin I work for Microsoft. Previously I was a partner at Eagle River, a private equity firm established by Craig McCaw. I am on the board of directors of Kymeta. Jan. 04, 2020 2 min read + Your Authors

1/ As long as there is customer acquisition cost (CAC) "free" isn't going away. CAC will disappear exactly never. What may go away is paying so much in CAC that 1) the impact on unit economics destroys shareholder value and 2) the business runs out of cash. 

2/ Finance businesses are notable for high CAC. This attachment from about a year ago was circulating on Twitter yesterday. That businesses will stop using loss leaders that have been in use at least since salty snacks were first given away in very early taverns is just wrong.

3/ Acquiring customers at scale economically is a challenging problem. That businesses will just use paid media and no longer rely on freemium isn't remotely credible. For example, free-to-play accounted for 80% of dollars spent on digital games in 2019. 

4/ LTV is a model. All models are wrong, but some are useful. People can use the model to justify CAC in ways that will kill the business by using broken assumptions. All assumptions should have a margin of safety. Whenever you see a spreadsheet, always focus on the assumptions.

5/ The Kelly Criterion is another example of a model than it useful but should be applied with a margin of safety. "We are not able to calculate exact probabilities [and] there are things that are going on that are not part of one’s knowledge." Ed Torp 

6/ Thorp: "Something like half Kelly is probably a prudent starting point. You might increase from there if you are more certain about the probabilities and decrease if you are less sure about the probabilities.”

Always look at assumptions carefully. 

7/ "Though many DCF models do incorporate sensitivity analysis (typically a grid of values...) ... investors should look to the value drivers—sales, margins, and investment needs—as sources of variant perception." Mauboussin. 

8/ "Even sensitivity analysis based on the value drivers is generally flawed because it fails to consider the interactivity between value drivers. Proper scenario analysis considers how changes in sales, costs, and investments lead to varying value driver outcomes." Mauboussin

9/ Here's a spreadsheet valuing an Amazon Prime customer. You can model the unit economics yourself (eg using this  ).

Which assumptions/model inputs produce the greatest sensitivity in the outcomes?

What should be done to reduce risk in this case?

10/ What happens in this spreadsheet measuring Amazon Prime unit economics when customer churn rises?

If this was disclosed by every company would it be easier to calculate NPV/LTV?

How many methods are used to calculate customer churn? Which method is most valuable?

You can follow @trengriffin.


Tip: mention @threader_app on a Twitter thread with the keyword “compile” to get a link to it.

Enjoy Threader? Sign up.

Since you’re here...

... we’re asking visitors like you to make a contribution to support this independent project. In these uncertain times, access to information is vital. Threader gets 1,000,000+ visits a month and our iOS Twitter client was featured as an App of the Day by Apple. Your financial support will help two developers to keep working on this app. Everyone’s contribution, big or small, is so valuable. Support Threader by becoming premium or by donating on PayPal. Thank you.

Follow Threader