What You Can Measure, You Can Manage
Measure and Compare Everything That Moves
A key element of our business development program is measurement. With each business development initiative we implement, we always measure and compare results. Always test one idea against another.
It is amazing how few companies do this. Few test any aspect of their marketing and then compare it to something else. They bet their destiny on arbitrary, subjective decisions and conjecture.
One way of saying it is like this:
We do not have the right or the power to predetermine what the marketplace wants — to do so is arrogant in the extreme.
Interesting thought, isn’t it? None of us want to be considered arrogant. And the way to avoid it is by measuring the results of different approaches. But it’s not just measuring — it’s measuring everything! Here’s a simple example: ever wondered why McDonald’s stores are red and yellow? (It’s not by accident!). They tested different approaches (colors in this case) and found red and yellow produced better results for them.
The point is — and this is not guesswork — when you test one approach against another and then carefully measure and tabulate the results, you will be amazed that one approach almost always produces substantially better results by a significant margin.
Measurement has a purpose. It’s to demand maximum performance from each marketing effort.
Any business CAN achieve immediate increases in sales and profits merely by measuring what’s currently happening and then measuring what happens when some element in the marketing mix is changed. Let’s say it again — what you can measure, you can manage.
You need to measure every variable in your business. Then measure what happens as various elements are changed (taking care to check that only one thing is changed at a time, of course).
But don’t stop. Keep measuring and testing to find out “how high is high!”
Keep experimenting to come up with even better approaches than the current “control” approach.
Your control is the particular way of doing things that proves to be the current best performer.
Until you establish your control, it’s impossible to maximize the marketing effort.
If you run advertisements in newspapers or magazines, test different approaches, different headlines, different hot-button emphases, different packages, different rationales, different pricing, and different bonuses on top of the basic offer.
Test positioning in the front, back, right, or left-hand side of the page. Test medium against medium; test the television or radio stations on which the commercials run.
Measure the number of responses, prospects, and resulting sales for each specific advertisement. Then compute the cost-per-prospect, the cost-per-sale, the average sale-per-prospect, the average conversion-per-prospect, and the average profit-per-sale against your control. This reveals the obvious winner — the control that you will keep running until a better control beats it.
In all of this, it’s important to remember the cost of things and the leverage you can get from them. For example, a salaried salesperson costs your business the same fixed amount, whether they make one sale a day, three sales a day, or four.
An advertisement costs you the same amount of space, production time, or airtime whether it produces 100 prospects, 1,000 prospects, or 10,000 prospects.
Therefore, it stands to reason that you should measure and test different approaches and find those that outperform all the others, then use those approaches to maximize the investment.
Measure everything starting right now.
Let’s look at some specific examples:
Testing prices is a major source of surprise for many people. Different prices on the same product often outperform one another by an enormous margin.
It’s not a matter of increasing prices — it’s a matter of testing them. And testing can be done imperceptibly.
One business owner was advised to increase his prices. The owner did not welcome the idea at all. Like most business people, he rejected the idea as ridiculous. “I’ll lose all my customers,” he said.
So the accountant devised an interesting test. “Let’s take the slow-moving items,” he said. “They represent about 30 percent of your $450,000 revenue. Let’s test increasing the price of those items by just 10 percent. And, of course, let’s measure the results.”
The result? No decrease in sales of slow-moving product lines at all, and an extra $13,500 on the bottom line.
Of course, the client would not have realized that gain had he not been game to test. Reluctance to test is something you’ll have to overcome — particularly when it comes to testing prices.
But deal with it by measurement. And remember, you’re not changing the prices forever. You’re simply testing.
Tattoo the phrase on your forehead — what you can measure, you can manage.
The truth is this — very few business owners measure anything, let alone test it. Instead, they rely on their own intuition about what is working and what isn’t. Ask yourself, “Do I know the cost of acquiring of a new customer, and do I know the lifetime value of that customer?”
We advocate that you should never do a single bit of marketing — not a single advertisement, mailing piece or whatever — without being able to measure its results.
For example, you should never place an advertisement without some kind of response mechanism. They should always include a coupon to return, or have a toll-free number to call.
Then for each placement of the advertisement where you’ve changed an element, have a different “key” number in the advertisement. Maybe even a different department to write to.
We suggest that you never stop at one advertisement. No matter how good the response, always test other approaches to see if one advertisement is producing even better results. You want to find out how high is high.
Once again, this means you must measure and then test one headline against another, test one price against another, and test one guarantee against another.
Whenever one advertisement demonstrably outperforms all others, that advertisement becomes your “control”. Then try to beat it. As soon as you find one that works better, that new one becomes the control.
However, here is an important caution. Make sure you never stop running an advertisement just because YOU have grown tired of it. Your judgment doesn’t count here. Again, the only valid judge is the marketplace. If an advertisement keeps producing, keep running it. Some successful mail-order advertisements have been running unchanged for 25 years. One presumes they’ve tested it and it works — they haven’t yet found a way to beat the control.
Headlines are the most important element in any advertisement — therefore it’s important that you know the effect that a different headline can have on the effectiveness of an advertisement.
You can conduct the test in a number of ways. First, if you are using display advertising, you can pre-test the headlines against one another in small display classified-type advertisements, or even in straight classified advertisements.
Next, if that’s impossible, you can test space advertisements by running the smallest test region available. Many magazines, newspapers, and other publications allow you to run in only a geographic fraction of their total circulation. You may pay a premium for this, but it’s much cheaper than paying the full price to run in a full national circulation.
Third, if that’s impossible, you can rent segments of the list that your piece will ultimately be disseminated to, and then mail 500 or 1,000 copies of Advertisement “A” and 500 or 1,000 of Advertisement “B”. One of the two will almost universally out-produce the other. After determining which advertisement pulled better, "roll it out" to a larger geographic area or larger number of names or total circulation, etc.
If you have salespeople, take one salesperson who is in a certain territory and another who has a comparable size or demographically balanced, equivalent territory. Have salesperson “A” present a package, product, offer, price, etc., in a certain way for a set period of time. Have salesperson “B” present the same package, product, offer, price, etc. - this time changing one variable.
At the end of the time period, analyze which approach produced the best results, and then integrate it throughout the system. After this, begin again with another test, expanded this time, to improve on that sales method.
Pre-validate the smallest test area first, and then go a little larger. You may go three or four times that size and validate it again. If it continues to work, keep expanding. Never “go for broke” unless you have to. In other words, if your client has fifteen salespeople and you test two, and one approach out-produces another by four times, you would have to deduce that the successful concept is four times more powerful.
But be careful. It could be that the salesperson who did well is just more articulate, more excited, and more conceptually able to embrace the benefits of the product. It may not be a better concept at all.
The only way you can really verify your results is to do the test again in another expanded test site. This time, use three salespeople to try it one way and three salespeople to try it the other way, and see if the results are still the same, or at least comparable.
Here’s an example to which we can all relate:
Take the simple (and regrettably, most frequently used) words you hear when you go shopping. The words are, of course, “Can I help you?” These four words are costing retailers (and others) a veritable fortune.
If these retail salespeople replaced these words with opening remarks like, “Hello, have you been in here before?” or “Hello, thank you for dropping in. It’s really nice to see you,” you’ll get totally different responses than the sales-killing “No thanks, I’m just looking.”
So, even simple words at this part of the sales process can make a huge difference. In fact, removing “Can I help you?” and replacing it with one of the phrases we outlined above can make a 16% difference to revenues and, on a $1 million base, that’s $160,000 annually by simply testing one tiny phrase!
Any time you’re planning to send a mailing to a large group of people (and large is a relative term), first rent the smallest representative segment you can. It may be 500, it may be 1,000, or it may be 5,000 names that are the most representatively balanced segment available.
In other words, if you live in a city with five different sections, and if you only mail to the most affluent or the most poverty-stricken, the results are naturally going to be dramatically skewed.
If the test mailing performs well, it is appropriate to either mail to the whole list or, if that’s too extensive, mail to only a slightly larger segment to verify or re-validate your results.
You can also test by segment to evaluate one approach or one concept over another. If you have only one concept that you are enthusiastic about, don’t spend a lot of money on it. First, for a small, modest amount of money, prove or validate that it works.
One great example of the value of testing is Steve Houghton, an Australian supplier of a personal alarm called “The Walkeasy Personal Alarm.”
Steve used radio advertising (advertising that he got on what’s called a “per inquiry” basis — more about that later) to test different approaches, different prices, different commercial lengths, different words, and even presenters of different sexes.
For example, he found that changing from a male voice to a female voice in the commercial doubled his sales. He found that moving from a 60-second commercial to a 45-second commercial, with a 15-second “tag” played later in the break, doubled his sales.
Steve made over $1 million in additional sales by exploring all the possibilities. In fact, when testing what price he should charge for the alarm, he discovered that the alarm he was barely selling at $9.95/unit sold like hot cakes at $39.95/unit!
Measuring and testing is absolutely necessary to maximize the potential of every aspect of any marketing program. You can (actually, you must) test at least the following:
Price, Offers, Packaging, Headlines, Words (Phone), Graphics, The List, Scripts, Colors, Guarantees, Placement, Copy, Timing, Advertisement Size, and Gender/Demographics
Be aware, though, that when you’re measuring the results of campaigns, it’s not enough to simply count the number of people who respond. The quality of the response is just as important. In other words, a lot of people may inquire, but how many of them buy?
In fact, it even goes beyond the number of people who buy a particular product. What you really want to measure is how much a customer is worth to the client in monetary values over his or her lifetime.
We call this figure the “Incremental Lifetime Value” (ILV) of the customer. It is also called Marginal Net Worth.
Less than one businessperson in a thousand really thinks in terms of incremental lifetime value, or knows how to calculate it. Yet, the calculation is easy and yields enormous benefits.
For one thing, if you know the incremental lifetime value of a customer, you can determine in advance how much you can afford to spend to acquire that customer in the first place. Moreover, you can reliably predict your cash flow well into the future.
For example, here’s a specific example involving the Mail Order Division of a Coffee Roaster. Let’s see how important it is to look at the ILV concept.
The Coffee Roaster runs an advertisement that costs $12,000. The ad invites potential customers to take up an offer of a free coffee-making machine that has a retail value of $51.95 when they buy a sampler selection of fresh roasted coffee blends. They invest $34.95 for the sampler pack.
Part of the deal is that the customers consider a ‘Til further notice home delivery service, whereby each month an order of their preference is sent to them and charged to their credit card.
The hard cost of supplying the free coffee-making equipment and the sampler pack leaves a net profit of $1 per response. When you factor in the $12,000 for the advertisement, most people would immediately conclude that 12,000 responses are needed just to break-even.
On this basis, most of our clients would probably immediately dismiss the idea.
But let’s look more carefully at the numbers.
When you know that for each person who finds the TFN system to be a convenient way to buy that the average annual gross profit is $245, you discover that the break-even response rate falls from 12,000 to 49 people. That is, the Coffee Roaster will recover the front-end investment within one year if just 49 people respond.
When you also discover (herein lies the critical importance of testing) that the average customer stays for at least three years and refers an average of .2 new customers, the numbers become even more interesting. The ILV of a customer from this promotion is worth at least $882.
The Coffee Roaster ran the ad, and 300 people responded to the offer. Of those, 100 said, “Thanks for the gift, but I don’t want any home delivery.” Another 100 said, “Thanks for the offer, but I’ll call you if I want any more.” But 100 said, “Thanks, I love this arrangement and I love your coffee – please continue to send it to me.”
In this situation, the Coffee Roaster just parlayed $12,000 into an $88,200 return over a three-year timeframe. Do that four times a year, and it’s a pretty neat way to create a return of $352,800 for a $48,000 investment.
This kind of information is invaluable. For one thing, if you know that each customer will ultimately bring in more than $882 in profit, you can justify spending more up front to bring in that customer. It would be interesting to test an offer where the customer didn’t pay anything up front because of the value that’s at stake.
Interestingly, you’ll probably find that the quality of responses, although lower in number, will be higher when there’s an up-front payment – but that’s for you to test. Don’t bet the business on the test, but take note of the numbers. Remember, what you can measure, you can manage.
In addition, we now can estimate the Coffee Roaster’s future profitability and cash flow with greater certainty. That’s because we now know that whenever he runs that same $12,000 advertisement, he’ll probably generate about $88,200 in profit over the subsequent three years.
Also, there’s another benefit, which is not inconsequential. If the Coffee Roaster ever wanted to sell his business, showing a prospective buyer an incremental lifetime value projection, like the one above, could go a long way towards securing that sale.
The lifetime value of a customer is a critically important factor in any business. Put simply, unless and until you know what it is, you have no idea of how much to invest in advertising to create a new customer.
You calculate Incremental Lifetime Value by knowing: a) the average length of time a customer stays, and b) the average net profit from sales that you make over the time your average customer continues to deal with the business.
Clearly, an important business strategy follows from this — the more you develop “back end” sales, the more valuable your customer base is because this will reflected in a higher ILV. In other words, you must nurture profitable customers.