Interesting exercise. I'd like to see the justification for the "3" on the right side of the equation though. Does anyone know where that came from or have a reference?

A lot of the estimates are really bad and completely ignore sensitivities due to assumptions, especially with CAC in the denominator which can have a massive effect when the range of ratios offered in the article is 0 < r < 3.

The difference between a LTV / CAC ratio of 2 and 3 is 50%. If CAC is assumed to be $3000 when the ratio is 2, then it only needs to drop by 33% to $2000 to get the ratio up to 3.

The author appears to consistently guess at CAC and then completely ignore the fact that a slight shift in the right direction would quickly put a company back in unicorn territory.

What's badly needed here is a sensitivity analysis.

For example, he lists ZocDoc's CAC as in the range $1k to $10k.

If it's $1k then their LTV/CAC = 4.8 (wow good!)

If it's $10k then their LTV/CAC = 0.48 (wow horrid!)

The author chose $3k, seemingly arbitrarily.

There's nothing wrong with arbitrary assumptions in general, but in this case a lot more scrutiny is needed before we judge if its a unicorn or donkey.

Interesting exercise. One can nitpick some of the math and assumptions, but the message is directionally correct.

The problem too many of these so called unicorns have is that there's not much in the way of real fundamentals supporting their lofty valuations. The valuation is largely based on hype. A lot of these companies valued at $1billion + may realistically only have a value of a few million based on their fundamentals.

In once sense that doesn't matter. If you're an early investor in such a unicorn you only need to convince some sucker to buy your shares while the hype is still hot. However, one needs to be a total fool to not understand that in such a game the music always stops playing at some point and someone is left holding a bag full of worthless $#&!.

The early players in this game have long since cashed out and are on a nice beach somewhere. The challenge for today's 'unicorns' is that they're starting to venture into nightmarish territory for businesses... i.e. massively overvalued without the fundamentals to stop a free fall. The number of recent tech IPOs where shares have plunged 50+% shortly after floating (or were forced into a recent big down round) are a strong sign that the market's tolerance for hype-based valuation is disappearing quickly. For those companies in that boat it's quickly going to become a game of put up or shut up. The solid business will survive but the rest will implode or be sold for pennies on the dollar in a fire sale down the line.

I think according to this analysis unicorns don't exists...

I haven't checked how the magical 3 is arrived at but assuming it's reasonable I think a more valuable exercise would have been to hold the 3 constant and plug in the values you're most sure about and then speculate about the missing variables. For ZocDoc with its 1-10k CAC range you could then speculate if they could reach the CAC that would lead to the magical >3.

So basically I think it's more valuable to use this as a tool to check if you think a startup can realistically optimize towards the >3 or if that is futile (see how they can influence LTV and CAC respectively) instead of trying to "rank" startups by plugging in LTV/CAC if you're not sure about them.

Interesting read either way.

I was a little questioning of why it was 3, but it isn't an unreasonable ratio after some thought: 1 would be borderline absurd, 2 could work but might not work with your operational costs, and 3 should represent a reasonable return on investment.

There is probably a closer, more precise number, but 3 is a decent number to use; if you're looking for businesses that generate a lot of revenue from customers, without spending a lot to acquire them or serve them; it works out.

Otherwise, great article and I will definitely be keeping this in mind in my own project. When I did the math it came out to 8.4, so I've got more wiggle room in the CAC than I thought.

Like others have mentioned, I'm very interested in how the author arrived at the ratio of '3'. At the very least this ratio doesn't take into account how capital intensive a company is, since fixed costs seem to be what's left (my understanding could be wrong here depending on which definition of 'margin' is being implied here).

Wouldn't this imply that a company with a lower ratio that is not very capital intensive would have a higher "corrected ratio" than a company which is much more capital intensive (or even simply has a higher cost of capital)?

Reminds me of this: http://pbfcomics.com/253/

So are we supposed to ignore that unicorns aren't real and that donkeys are useful pack animals?

I mean, say I show up on a farm looking for work and you tell me there is a cart of manure to haul somewhere, and then conditionally tell me that either A) I'll have a donkey to do the job, or B) I'll have a unicorn. Option B sounds more like I'm doing the job by my lonesome.

The ratio of 3 may or may not be correct, but the specific analyses here are absolutely worthless as at least one of the model inputs for each example is just pulled out of thin air. He could be out by an order of magnitude.

What if the service has potential to drive LTV, but intentionally refuses to monetize in order to increase audience satisfaction and growth?

unless I'm misunderstanding, isn't this double counting the costs? wouldn't the margin (revenue-costs / revenue) already take into account the cost of acquiring customers?

in their example, 52% margin and a $400 CAC:

  LTV/CAC = 0.25 years x $2160/year x 52% / $400 = 0.70x

Great article. Keen to see how it applies to Xero and Shopify. I'll bet they are donkeys too.

i'd be interest in his analysis of popular YC companies, AirBNB for example