There has been a lot of good stuff written over the years on the topic of calculating customer lifetime value (LTV). Thus, it amazes me how many times I discover faulty thinking when I talk to entrepreneurs regarding their LTV math. One portfolio company executive confessed to me last week that he knows he is doing it wrong but he just didn't have the time to research the best way to do the LTV calculation.

Since I see a few common patterns of mistakes, I thought I'd add to the LTV literature and point out the top three reasons many investors roll their eyes when they see entrepreneurs present inflated, poorly constructed LTVs:

**1) Your churn rate is understated**

One important component to an LTV calculation is the churn rate or cancellation rate. Many blogs suggest you simply divide 1 by your monthly churn rate to get to a number of months of duration that you can expect to collect revenues from your customer. Thus, if your average monthly churn rate is "c", the number of months of revenue you will receive over the lifetime of a customer is 1/c.

The problem is that many early-stage companies have no idea what their average, long-term churn rate really is because they are simply too young. When they have 6 month or 12 month or even 18 month cohorts, they extrapolate from those cohorts and come up with an absurd time period for their customers to stick around generating revenue. For example, if you have a 2% monthly churn rate in your first year, then some folks will extrapolate their monthly revenues out 50 months. A monthly churn rate of 1%? Then multiply that monthly revenue by 100.

As Jason Cohen points out, it's just not realistic that in a wildly competitive, dynamic technology market, a company can expect to hold on to its customer *on average* for 8-10 years. And, in my experience, you are so hyper-focused on satisfying and servicing your early customers that extrapolating your early churn rate just isn't going to be accurate.

To fix this potential issue, I recommend you pick a fixed cap number of months – conservatively 36, or three years – and recalibrate your LTV math accordingly. Your new expected months of revenue (N) would now = [1-(1-c)^36]/c. For example, if your churn rate is 1%/month, instead of assuming 100 months of revenue, you calculate 30 months. Anything beyond 36 months just doesn't seem credible – and shouldn't even matter that much when you think about the next issue – a start-up's cost of capital.

**2) Your cost of capital is too low**

Ask an entrepreneur about their cost of capital and you'll likely get a blank stare. Cost of capital is the rate of return that an investor who provides capital expects from investing that capital. Today, the United States government has a cost of capital of nearly zero – for example, it can borrow money for 10 years and pay only 2% interest. 2% per year is the expected return that an investor in US treasuries requires because the risk of holding an IOU from the US government is so low.

For a start-up to raise capital, it must sell equity to venture capitalists or other investors that expect an annual return more like 30-40% in exchange for the high risk that the company will never be able to pay back the investor and the investment will be written down to zero. Thus, the cost of capital for a start-up (and the dilution a founder faces in exchange for that capital) is very high. Therefore, back end loaded cash flows are not nearly as valuable for a start-up as front end loaded cash flows.

That's a bit of context as to why start-ups need to highly discount future cash flows when calculating their LTV. I suggest 3%/month which results in a roughly 30% annual cost of capital. Thus, if you are receiving $100 in recurring revenue, you should value next month's $100 in revenue as $97 and month 2 as $94. In practice, combining this point with the one above, take your number of months of revenue (say, 30) and use the 3%/month discount rate to calculate the value of the months of revenue = [1-(1-3%)^30]/3% = 20 months of revenue – 1/5th what you would have calculated if you had simply used 1%/month churn rate with no time limit and no discount rate!

**3) You forgot about Gross Margin – and you're probably overstating them.**

I recently received a board deck from one of my portfolio companies which treated revenue as the numerator in their LTV calculation. Entrepreneurs sometimes forget that a dollar of revenue isn't worth a dollar in incremental contribution. Instead, there is real cost to produce this revenue: a cost of service, processing, data, storage, media, overhead whatever.

Many early stage companies don't yet have experienced CFOs who can help them with precise gross margin calculations, so they assume a gross margin that is too high. SaaS companies think "mature SaaS company margins are 80%" so I'll just use that. But you are not mature. Your executive team spends more time selling and servicing than you account for. Your engineers spend more time servicing customers over time and addressing issues and bugs and feature requests than you account for. Thus, your COGS (cost of goods) are understated and your gross margin is overstated. Salesforce.com has a gross margin of 75% with their scale of $6 billion in annual revenue. Can yours really be the same or even 5-10 percentage points better? And are you sure your gross margin calculation is factoring in all variable costs not related to customer acquisition or are some costs sneaking "below the line" into, say, SG&A?

To fix this one, the rule of thumb I suggest you use is to discount an additional 10% points beyond whatever your finance head says your gross margin is. Thus, if you think your gross margin is 70%, assume for LTV calculation purposes 60%. So, in the example above, instead of summing up $100 revenue over 20 months (factoring in a shorter time horizon and a higher cost of capital), you would sum up $60 over 20 months. Add all three factors together, and instead of multiplying $100 in monthly revenue by 100x for an LTV of $10,000, you would be multiplying $60 in monthly contribution margin by 20x for an LTV of $1,200.

Conclusion

All of these factors – time realism, appropriate cost of capital and accurate gross margins – discount your LTV as compared to simpler methods. Sorry, but that is the reality of LTV math. If you have a business with strong network effects, there can be a reason to believe that your metrics will meaningfully improve over time. But another reality of LTV math is that absent strong network effects or other large benefits of scale, many times your metrics get worse with scale. I cover this phenomenon in another blog post and so will simply say: make sure you don't overstate early metrics with rosy extrapolations.

A mentor of mine is fond of saying that every business plan contains the same word in relation to its forecasts: "conservative". It is better to be truly conservative – or, dare I say, accurate – rather than letting a savvy, cynical investor do it for you.