Estimating The Value Of A Customer And Advertising Costs
Each customer has a value to a business. Some businesses seek to estimate this value with some precision. Other companies just do their best to satisfy customers and neglect all of what follows below. Knowing the value of a customer will allow you to calculate the "real" breakeven cost of doing a promotion. Just knowing how to calculate the value of a customer will make you a better businessperson.
The value of a customer is a relatively simple concept. The value of a customer is just the total profit that an average, initial customer generates for a business. Suppose you only offer one product for sale and only sell it once to any given customer. Then, the value of the customer is the profit from that sale alone.
However, most successful businesses don't sell once to a customer. They generate repeat business. The true value of a customer must account for profits from repeat sales made after the initial purchase. This concept uses the time value of money just a bit. Unless you are predicting long customer relationships, the inherent uncertainties of the estimates will probably dominate the effect of properly compensating for the time value of money, but we will show the full and proper calculation.
Let's take the example of a fictitious company called "ABC Web Hosts." (Note: The costs used below are not meant to be representative of real-life web hosting costs per se. The goal is only to show you the methodology of calculating customer value.)
Suppose to sign up a new account ABC Web Hosts charges $25 and $5 of that is expense to set up the new account. So, it looks like we earn $20. And, that is true for each new customer, neglecting costs to acquire the customer, i.e., marketing costs. Also, each person pays $10/month for hosting, and assume it costs $3/month to administer the account. So each month of hosting purchases generates $7 in profits.
People sign up for:
1 month Profits = $20 + 7 = $27
2 months Profits = $20 + 14 = $34
3 months Profits = $20 + 21 = $41
Suppose 1/3 of the people sign up for each of 1, 2, and 3 months. The total profits on average are
($27 + $34 + $41)/3 = $34 per customer
which corresponds to the value of a customer who buys two months service. (The actual distribution of customer purchases would need to be based upon ABC's actual business results. Note: We neglect those paying for more than 3 months for convenience. Generalizing the calculation by adding 4 months, 5 months, Ö is trivial.)(Notice 2 months corresponds to what I refer to as "Customer Life Expectancy" in Thinking Like An Entrepreneur.)
Now, the question is: "What is the overall value of a customer?"
It depends upon the renewal value of a customer. (Note: These calculations are most appropriate to newsletters, magazine subscriptions, mail-order businesses, ongoing services, etc, where initial buyers are converted into repeat buyers.)
Suppose that: 50% of the initial customers renew for 3 more months
20% of the initial customers renew for 1 more month
30% of the initial customers don't renew
(Again, we are just taking a simplified example. Some would renew for six months, etc. The actual distribution is best predicted from historical data for the actual company in question. Obviously, renewal rates vary, with quality of service and product being a strong factor.)
The average renewal length is (.5)(3) + (.2)(1) + (.3)(0) = 1.7 months.
Now neglecting second, subsequent renewals, the first renewal value of a customer is:
(.5)($21) + (.2)($7) + (.3)(0) = $11.90
Or thought of another way (1.7 months)($7/month) = $11.90
Notice two important things. One, we were careful not to incorrectly add to the ongoing profits the one-time set-up profit of $20. That doesn't apply upon renewal. Second, we are neglecting the time value of money, because the time period in question is short (<6 months ). Recall, the average customer has been with us 2 months before he renews, so formally we could discount this $11.90 back a couple of months. Flip it. It is not material.
We neglect the time value of money for short periods of time.
So the average customer life value = $34 + $11.90 = $45.90. (Neglecting second time renewals which will follow.)
Using the above value for the average customer value would be a very conservative estimate. It neglects the effects of those customers who renew a second time. The further into the future we get, the more skeptical we must be of our ability to accurately predict customer valuation.
Let's assume that 60% of all those who have renewed once, renew for a full one more year. (Again a simplifying assumption. You would use historical data as your best estimate of second renewal conversions.)
(0.7)($7)(12)(0.6) + (0)(0.4) = $35.28 (second time renewal value) would be added to the value of a customer ($7 profit per month times one year). Notice that long-term customers add significantly to the average value of a customer. This second renewal amount value of $35.28 is comparable to the $45.90 which measures customer value through the first renewal only.
Notice the all-important factor of 0.7 which accounts for the fact that only 70% of the initial customers have renewed the first time (the sum of 50% renewing 3 months and 20% renewing one month). This is easy to miss. We are interested in the value of acquiring a new customer and must properly account for customer attrition. If a customer hasn't renewed the first time, obviously, he can't renew a second time. And, again, we could properly discount this $35.28 by 3.7 months, as it is received 3.7 months from the first order of the customer. The effect here is small so we neglect doing this.
Lesson: To succeed in many businesses, you want long-term customer lifetimes.
Suppose all those who renew a second time are satisfied and stay for 3 more years in addition to the above time, i.e., they renew a third time. That would add (0.6)(0.7)($7)(12)(3) =$105.84 to the value of a customer. (If only 90% renewed the third time, we would pick up a factor of (0.9) to account for the further attrition. A customer once lost is not regained. At least not in our example! We must carry the factor of 0.7 along with our calculations. Think of it this way: Only 70% of the initial customers are around upon first renewal, and only 60% of those are around on second renewal are viable third renewal candidates. It was assumed all these hardy folks are now staying with us for the 3 more years. (This is the third renewal value)
Notice customers who renew a few times probably like what the company offers and you should have relatively high repeat business from them if your company is good. You will have some loss however. But, subsequent renewal rates beyond the second are usually higher than first time renewal rates in many areas like newsletter publishing.
Now 3 years is getting to be a longer period of time, and we should examine the time value of money. The typical renewing customer has been with us now for about 15.7 months (2 months + 1.7 months + 12 months.) Assume the customer who stays with us for the three years renews annually. (As usual, a company's actual historical data would be the basis for the real numbers. Don't confuse customers staying on for 3 years with paying for three years in advance! How they pay needs to be considered also! Our simplifying assumption is that they pay for one year service in advance, but stay on three years in addition to the 15.7 months.)
Using the time value of money, we could discount each year's value back to a given point in time. (Properly, we should really discount it all the way back to the very start of acquiring a customer. I don't want to do that as it forces us to break time value of money into monthly periods. This is covered in Thinking Like An Entrepreneur and many other books, and I assume you can generalize time value of money calculations to a monthly basis.)
For example, at a discount rate of 10% and discounting the value only back to the time of the third renewal (15.7 months), each annual payment contributes as follows:
(0.7)(0.6) (($7)(12) + ($7)(12)(1/1.10) +($7)(12)(1/1.10)2 = $84 + $76.36 + $69.42 = $229.78)) = $96.51
Recall just taking (0.6)(0.7)($7)(12)(3) yields (0.6)(0.7)$252 = $105.84 which is a bit too high. The difference is about 9% between no time value of money and time value accounted for.
I wrote the above in such a way that we donít forget the attrition factors of (0.7)(0.6)
All rightty then. Where are we? OK. We must calculate the overall value of the initial customer.
It is: $34 + $11.90 + $35.28 + $105.84 = $187.02 where we have been just a wee bit sloppy with the time value of money on the last term of $105.84. And a bit wee less sloppy with the middle terms! To improve the calculation we could use $96.51 as the last term. Then, we could more properly discount the last term of $96.51 back one more year. Or, we could break the calculation into monthly time value of money periods and do the whole thing entirely properly (Neglecting the small difference between 1.7 and 2 months).
But, that is really unnecessary.
Lesson: The uncertainty of business conditions and other factors are large compared to the time value of money over short periods of time. Inherently, the validity of any such customer life value calculations is less certain with time. Yet, you should probably account for the time value of money. I leave it as an exercise to put the above onto a monthly footing and account for the time value of money in full detail (i.e., I should have chosen an example where everything is renewed yearly to avoid distracting complications I didn't want to go into!).
Next up. How much should you pay for a customer? Let's suppose you do the above in brutal detail, account for the time value of money and get an average customer value of, oh, let's just say about $170 when discounted to the present (I discounted the improved last term of $96.51 back one more year. And, we neglect sales beyond about four years, which is getting pretty far into the future).
Some mail order books say it is ok to spend right up to the value of a customer on your advertising. (This can lead to what I briefly call a "revenue trap" business, one that grows revenues significantly, but never generates decent profits. As mentioned in Thinking Like An Entrepreneur this is a big issue when valuing a mail-order business. You need to be aware of this before you blindly start applying Price-To-Sales ratios for company valuation.) The discount rate is not enough to compensate for the risks unless you set it high enough.
We used 10%, a rate of return you could get investing passively in quality common stocks. If you were to acquire customers through advertising, you would be aiming for the discount rate of return, which is inadequate given all the risks, if you acquired customers right up to a cost of $170.
Suppose a year from now you received $110. How much would you be willing to pay at present for this? Well, if you pay $100, you have just attained the "market" discount rate of return of 10%. If you know you could get $110 in a year with higher safety, why would you invest $100 for a higher-risk endeavor with only the same return? That is what you are doing with the $170 if you are willing to spend the full $170 to acquire a customer. You are just breaking even when you account for the time value of money. That is not good enough.
Lesson: Either you must allow a sufficiently generous discount rate comparable to the high risk of return you demand for the investment in the advertising promotion, or else, you must buy a customer for less than the estimated customer value, if you choose to use a conservative discount rate.
Example: We might use 15-20% as a "safer" more appropriately compensating rate of return. Or, let's just, at most, pay 50% of the value of a customer. Compare this to buying undervalued common stocks. If you are not demanding enough and pay $100 a share for a company worth $100 a share, it is difficult to make a lot of money.
You want to buy a company for a strong discount to its "intrinsic" value. In the same way, you want to buy customers for less than they are really "worth." Conversely, if you are too demanding and greedy, saying you will only pay, at most, $1 a share for a company whose value is "really" $100 per share, you will never have the opportunity to buy the stock. You probably will never see that price.
Similarly, if you demand a profit from your first promotional advertising without accounting for renewals and follow up sales (in our example, this is the initial order for 2 months' hosting generating a profit of $34 before allowing for advertising. So "too greedy" is defined as demanding an advertising cost per new customer of less than $34), you probably won't find any "viable" advertising outlets.
A reasonable target might be to acquire a customer for half his worth or less. Here, about $85. From a profitability standpoint this is good and it reduces our risk. But, be sure to examine the cost from a cash-flow standpoint. Remember, if you are paying up-front for a customer, most of whose worth will be returned to you after the first, second, or third year, you must be alert to the cash-flow consequences. (Basic cash flow is covered in Thinking Like An Entrepreneur)
We assume we collect $34 right away. We get $11.90 a couple of months later. Then, we get $35.28 about 3.7 months after acquiring the customer. Then the next payment is not until a year later. So our free cash flow is only about $81.18 within the first year, occurring within the first half of the year. It is "safest" if we try to acquire a new customer for no more than about $80 per customer.
Now to bring all this to conclusion, we assume we are willing to pay $80 to acquire a new customer. The customer's value is about $170. Suppose you buy banner advertising for $10 per 1000 exposures ($10 CPM, cost per thousand). If we get a 1% clickthrough rate of these people going to our site from the banner and if 1% of the people visiting our site become customers, we have achieved one customer for 10,000 exposures.
(10,000)(0.01) =100 people coming to our site from the banner
(100)(0.01) = 1 customer acquired.
At $10 CPM, 10,000 exposures cost $100. This is a bit outside of our desired cost to acquire a customer. If we can live with the cash-flow consequences, we might choose to run the advertising campaign.
Suppose you can increase both clickthrough and conversion to 1.1%. Then for each 10,000 exposures, we obtain 110 people visiting our site, and, of these, we acquire 1.21 customers. This is a cost per customer of $100/1.21 = $82.64 which is much closer to our target. If we could lower the CPM, that would help also.
Lesson: Increases in clickthrough (number of people seeing our banner who go to our site) or conversions of people visiting our site to buying customers have significant impact upon profitability of promotions. For example, in web hosting, maybe the costs per servicing an account drop sharply with more customers. This would affect the value of each customer via the changing expense which we just set at $3 per month. Also, customer happiness and satisfaction is crucial to having profitable customers.
This is all tenuous, as expenses change, competition might offer better deals, etc. Your estimates of customer renewal are tenuous and estimates only. Buying customers for up-front amounts approaching the customer value has cash-flow consequences. Buying customers is like buying stocks. You need to acquire them at a significant discount to their value to really succeed.
Finally, notice where the non-math enters the picture. The artistic, the psychological, the creative, if you will. It is in the customer service. With bad service and products, you won't have renewing customers. You will have one-time buyers. It is in marketing. How many people clickthrough, for example, is a function of banner effectiveness and design. (Notice: We did not allow for any extra marketing expense for conversion of a first-time customer to a repeat buyer. We assume an e-mail is all it takes. In other businesses, such as newsletters, a certain marketing amount is spent upon generating the repeat subscriptions. The cost needed to get renewals would need to be considered.
Learning how to calculate customer value gives the marketer the ability to better evaluate whether or not a particular advertising campaign is worthwhile.
Those who have read Thinking Like An Entrepreneur might want to reread the chapter on expectation values and see how closely the concepts there overlap with customer valuation.