I do a lot of writing and speaking, but the best parts of those activities are the people I meet. People who engage with me to teach me something. A few weeks ago, I appeared with Alan Rimm-Kaufman at an AMA Hot Topics in Search Marketing event. In my talk, I casually mentioned something I often discuss—that paid search bid management tools don’t optimize profit, but merely optimize profit margin (or worse, return on advertising spending or cost per action). Alan noted that remark and has spent the last few weeks engaging me on what I am missing. I thought it would be great to tell a story of how I came to think about campaign optimization, and also show you Alan’s comments on the story. Let’s get started with our tale of campaign optimization.
The replacement parts group within a large, well-known company (we’ll call them SlowCo) was struggling to optimize its return on investment. The parts group had traditionally been a place where the company made easy money. The not-so-brilliant executives were parked there, because it was thought they could do little damage. The less-than-stellar salespeople were dumped there. No matter how little effort and intellect was spent on the parts business, SlowCo still made money hand over fist, charging high prices for slow delivery, because their captive customers knew no one else to buy from.
Then the Internet came along and this sleepy business had to either wake up or go to sleep permanently. Competitors cropped up everywhere, undercutting the company with lower prices and speedier delivery. SlowCo parts customers disappeared in droves.
After a few months of outright losses, finally SlowCo’s management faced some hard decisions. The deadwood employees were weeded out, and new management decided to “take back our parts business” by focusing on lowering “cost per action”—they aggressively shifted sales that were being consummated over the phone to the Internet and they provided incentives to discount larger orders so that the costs of attracting orders went down. Unfortunately, it wasn’t just cost per action that was dropping—revenue kept dropping, too.
Alan’s comments: I hear a small warning bell by this point in the story already. It often isn’t a great idea to force a channel, and even more troubling to assume that customer value and service cost by channel will be the same forced and unforced. It is confusing causation and correlation. A few years back, the pundits were banging the drum that “most valuable customers shop in the store and online”—let’s assume that this was true. Therefore, “if we give coupons to store customers to use the Web, they’ll become most valuable customers.” Each summer temperatures rise along with ice cream consumption, but it isn’t the cones driving the mercury.
So the management team then focused on an all-out campaign to improve revenue. “Take back our business” turned into “Take back the Internet” as an aggressive paid search campaign ensued to divert customers shopping the new competitors so that they returned to SlowCo for their parts needs. They also increased discounting even further.
Alan’s comments: Discounts are an addictive drug. I probably sound like a sanctimonious preachy jerk, but it’s true. I think offers need to make rational sense. If you bought more red widgets than you can sell, it makes sense to reduce your prices and tell your customers they’re on clearance, admitting in effect you bought the wrong color or bought too many. No worries. But if you simply cut prices for no reason (perhaps due to a made-up PR day), you’re asking for trouble.
Last week, Bill Bass (formerly of Land’s End, now with Fair Indigo) dissed the idea of offering “CyberMonday discounts” just because everyone else does—hurrah to Bill for talking sense loudly.
Personally, I know of several big companies who even use addiction language when discussing their offer strategy: “We just can’t stop, the revenue feels so good.” This company has paid millions training their customers never to buy at full price. Ouch.
But the biggest change was that SlowCo added aggressive shipping discounts to match the competition’s fast delivery, and they promised to waive shipping charges for any part that failed to arrive when promised. Revenue jumped for the first time in several years, but when they added up the numbers at the end of the quarter, they had taken the biggest loss ever. Costs (especially shipping and rebated shipping costs) had so outstripped revenue that the parts division had become a drain on company performance.
Alan’s comments: Not to be snooty, but it took SlowCo’s managers three months to get around to looking at their P&L? These are direct costs, it sounds like, not corporate transfer costs—you’d think they’d see this faster. Apologies for snootiness, but what took so long? Mike’s comment: They aren’t known as SlowCo for nothing.
A new management team was brought in that demanded a return to profitability. All pricing, including discounts and shipping, were rigorously scrutinized so that profit margin could be calculated for each order.
Alan’s comments: Hooray. Our story takes a happy turn. I’m rooting for the new guys.
SlowCo’s new team made some dramatic about-faces. Paid search was dropped, as was the rebates for late shipments.
Alan’s comments: Paid search works for many marketers in many situations, but not all. In some verticals, the cost you must pay for each click are way above rational levels. I know someone in the inkjet refill business where the cost of entry is so low that he’s always bidding against newcomers who pay more than each click is worth, driving him off the first page. “They” (collectively) never learn, because when one competitor goes bust, another newbie crops up, so the paid search auction is never rational. The situation could be the same for these parts guys, maybe. Another place where paid search fails is in new product introductions (if something is novel, nobody is searching for it), but replacement parts don’t sound all that novel.
After these big changes, profit margins zoomed, but revenue cratered, as customers went back to buying from the competitors. Even though the accountants could show a profit on each item shipped, revenues dropped so fast that their fixed costs created yet another loss.
Alan’s comments: Problems for our brave heroes. Too many retailers focus on profit margin (a percentage) instead margin dollars (real money), or focus on unit contribution (without loading on overhead) instead of total contribution (with fixed costs appropriately allocated). At day’s end, you don’t deposit percentages, you deposit dollars. And the overhead needs to go somewhere. Once in a while, new marketing efforts or customer efforts can be excused from covering overhead, but that big nut—salaries, building, insurance, corporate—needs to covered somewhere. OK, now I am worried about the new management team at SlowCo, and don’t see a happy ending to this story heading our way.
As in any big company, this is the moment when the long knives come out. Some within SlowCo were calling for management to sell off the parts business to one of the new competitors that had cropped up so that they could get this albatross off their books. Chastened by their series of gaffes, the new management team finally stepped back from the situation to assess the real value of the parts business.
SlowCo spent several weeks studying the purchasing behavior of customers who had stayed loyal to the company’s parts business versus those that had not. What they found changed everyone’s outlook. Those customers that had stayed with the company’s parts business were the most loyal repurchasers of the company’s flagship products—the expensive high-margin products that the parts business supplied replacement parts for. That was where SlowCo’s bread was really buttered—parts were truly a sidelight, but an important one to SlowCo’s best customers.
So, the SlowCo management team dug deeper, eventually analyzing Lifetime Value to see whether customers with higher Lifetime Value purchased parts from them or their competitors. That study confirmed that the highest value customers purchased parts from SlowCo. What’s more, the constant changes in policies and pricing had done nothing but annoy this core group. It was the more price-sensitive, less-loyal customers that moved back and forth from the competitors, driving revenue up and down.
In the end, SlowCo decided to outsource their parts distribution to a fulfillment company known for low costs and speedy, reliable delivery, and they lowered their costs so that they were not much higher than their competitors. Revenue went back up, and they found that they were getting even higher repurchase rates from their parts customers for their main products.
Alan’s comments: Great they had good customer level data and were able to pull these insights from it.
Then SlowCo decided to experiment. They began offering replacement parts maintenance contracts as part of the purchase of their main products. They offered these contracts at subsistence pricing, with low or no profit. When they did, they found that their customers began choosing this low-cost maintenance in droves, driving revenue for parts to the sky, and raising total parts profit even though the profit margin was small. More importantly, they discovered that the customers who bought maintenance offerings became part of that core group of high-value customers who were far more likely to repurchase the main product when it was time.
Alan’s comments: Sounds like the contract offered real value to their customers. A known cost insurance policy, less hassle. And here it sounds like they there were able to show causation, not just correlation.
So, the SlowCo parts business, which started as a poorly-managed separate business unit trying to optimize its own return, had turned it around. Despite a series of missteps, they eventually came to their senses and looked at the contribution of the parts business to the company as a whole, which made it easy to subjugate the return on investment on parts to the bigger company ROI. Then, to everyone’s surprise, it turned out that the parts business was actually a critical driver in increasing customer loyalty, and an important driver to overall profitability.
Alan’s comments: Cue the happy ending music. I’m glad SlowCo realized the strategic importance of parts. It sounds like they got some good help ripping through their customer data. As with all stories, we should try to extract the moral. (“Avoid door-to-door crones offering apples.”) I think it is unrealistic to expect an automated bidding technology to set broader business policy (“parts biz good” vs. “parts biz bad”). Few if any direct marketers spread fixed costs onto their marketing programs. They don’t ignore fixed costs—that would quickly kill their businesses—they just don’t try to “variableize” them (hmmm… not sure if that is a real word!). But good PPC bid tools do focus on both sales quality and sales quantity.
Pardon a digression into high-school calculus, but when you’re maximizing a function, the constants don’t matter: the value x* that maximizes f(x) also maximizes f(x)-c, where c is a constant. That is, the optimal advertising level which maximizes marketing contribution (marketing contribution defined as sales less cost less advertising less variable costs) will also maximize EBITA aka profit (profit defined here as sales less cost less advertising less variable costs less fixed costs). Now, your total profit level depends critically on your fixed costs, sure, but the best advertising intensity doesn’t. That’s why direct marketers can optimize campaigns on contribution, and be doing the smart thing for profit.
So how do fixed costs come into the picture? If the retailer can’t generate enough contribution to cover fixed costs—even at the best advertising level—then he or she doesn’t have a sustainable business, and needs to get a new approach fast.
Back to PPC bidding, we’ve presented the basic economics of PPC bidding this summer in a post called Computing Optimal Pay-Per-Click Bids In 19 Easy Steps. (Hint: our true secret sauce—the fancy stats that really make bidding work well across the head and the tail—is mentioned in passing in Note 13.) Rereading that post now, I see that Note 19 mentioned fixed costs, and notes direct marketers usually do not allocate them, but critically do account for them.
Thanks Alan for your insightful comments in this month’s newsletter. By the way, if you’re not reading Alan’s RKG blog, why not? He writes for it and so do other folks from his company, the Rimm-Kaufman Group, which offers paid search and Web effectiveness consulting. And if you’re looking for a few other recommendations on blogs you should be reading, check out the Biznology blogroll.
