Averages Can Kill Your SEM Campaign

Averages allow you to take large quantities of rather complex data and make that data palatable and understandable. We all use them. However, averages are extremely dangerous in marketing. Think about what would happen if a state trooper pulled you over for doing 95 mph in a 55 zone, and you told him your average speed since you left the house was 52 mph. Similarly, your doctor doesn’t care what your last month’s average body temperature was if your current temperature is 109 degrees Fahrenheit.

When planning and executing your business strategy, media plan, and campaign optimization, averages can be deadly to your business. Often, averages in marketing analytics and campaign reports fail to expose critical opportunities. Averages can also hide significant spending waste — both in media and other areas of a business. If used without an understanding of their proper purpose, averages can be irrelevant and, worse, quite risky, because they are frequently misleading.

Some marketing and media practitioners understand the importance of looking beyond the averages, particularly if profit maximization is the true objective (and it almost always should be). For example, Jim Novo spends nearly every waking hour drilling down through data to discover the truth behind the averages. But, many search marketers are given maximum target cost-per-order/cost-per-acquisition (CPO/CPA) or return-on-advertising-spending (ROAS) objectives, and they’re managing campaigns based on average CPO/CPA or ROAS.

For most businesses, managing by an average is not optimal. Campaigns should be optimized based on the margin, meaning no medium should knowingly be purchased if it exceeds an objective based on true return on investment (ROI). After all, we’re in the business of profit maximization, or eventually we’re out of business.

Digging for Gold and Coal

Let’s use the simple example of a cell phone service provider. The marketing VP has directed the marketing team to use search engine marketing (SEM) with the stipulation new accounts be obtained for $100 or less. For accounts costing more than that, the profit on a customer is low or even negative.

After a two-month campaign, the team is ecstatic because the average CPO is $95, under the target CPO. Then the VP asks the marketing team to look at the numbers behind that average. The marketing team had tagged each listing individually, so could run a CPO report by listing and spending.

The report was sorted by keyword listing CPO, from most to least efficient. For the 400 keywords in the campaign, the most efficient 100 (first quartile) have an average CPO of $65. Wow, these are ultra-efficient, high-ROI gems! The second quartile (next 100 most efficient) has an average of $85, the next an average of $105, and the final batch an average of $125. The bad news hit hard. For all the segments of the campaign that resulted in a CPO over $100, the company has exceeded its maximum allowable CPO and very likely lost money. Uh-oh…

The average CPO number hid the fact much of the campaign was actually wasted budget. All the listings for which the CPO was over the max should have been redeployed in media that could deliver the customer acquisitions for less than $100. By redeploying that money, the company is better served. The marketing team should not be deceived by the averages, or think they are meeting their true objectives, just because the average of the campaign is being met.

An easy way to conceptualize the deceiving nature of averages is to imagine yourself getting paid the $100 per order and having to buy those orders yourself in the market for prices ranging from $25 to $150. You might buy the $99 order and make a dollar if the costs and time in consummating the transaction were low, but there is no way you would pay more than $100.

Think about the search results you have noticed for many keywords. You may have seen the experts in action, using granular data and spreads to make money. Both performance-based search marketing agencies and affiliates engage in something called arbitrage. For internal campaigns they set a maximum CPO below what they’re paid by marketers, then grab as much inventory as they can, up to their internal maximum. In the interests of additional volume, they might even allow the internal CPO to get close to the marketers’ actual payouts if they’re sure of their techniques and strategy. Affiliates have become very good at this strategy. I’ll cover the crossover between affiliate marketing and search marketing in a future column, but this spread is part of what drives the popularity of search marketing among affiliates.

Now let’s examine the marketing VP’s original cost of customer acquisition target. The VP specified a CPO/CPA of $100. This target was likely arrived at when the CFO pulled a report on average earnings per customer, and the VP and CFO determined a good percentage of earnings to allocate to customer acquisition. Because of how the target was determined, this provides the marketing VP with an opportunity to show just how smart the marketing team is.

After reading about lifetime value, the Pareto Effect, and acquisition costs on ClickZ, the marketing department has the technology team capture the listing IDs used in the search marketing campaign and record those IDs on each customer record during the sign-up process.

Looking back over a one-year period, the team finds the most profitable customers — who spend three times the average on the most comprehensive cell plans and rarely cancel — are more likely to come from certain keyword listings. Because of the profitability of those customers, the marketing VP and the CFO know spending $150 to acquire those customers would be a bargain. Similarly, customers who take cheap plans but don’t pay their bills and cancel early tend to originate from a different set of keywords and engines. These customers are not even worth $100; at most they’re worth $75. Now the search marketing team has broken its search campaign into subcampaigns, each with a different CPO target. These folks are now getting more of the right customers and fewer of the wrong ones.

Large differentials in customer profitability occur in almost every business. Online retailers make more gross immediate profit (both as a percentage and in total) on some products. Some product buyers tend to have different lifetime values based on purchase behavior or even returns. Travel companies make more on certain bookings or seek to capture high-frequency travelers. Insurance marketers want more high-commission policies with long renewals. Smart search marketers use all the information at their disposal to tune a search marketing campaign far beyond the averages.

Averages might give us a pulse of the campaign, but relying on averages alone or even telling your team to manage to averages will result in large missed opportunities and wasted budgets. Unleash the power of intelligent strategy and enlightened use of data to reveal the rich information that exists beyond the averages and execute a campaign that really maximizes returns and profit. Do it, and you’ll be way ahead of the competition.

Meet Kevin at the Jupiter ClickZ Advertising Forum in New York City on July 30 and 31.