Harvesting habits: How marketers can use purchasing cycles to increase sales

Even those who love a certain food — such as yogurt or frankfurters — often turn on their favorites and do not purchase the products for periods of time. Assistant Marketing Professor Sungho Park and his research partner studied these purchasing cycles, and they uncovered a valuable fact: timing promotions based on buying cycles makes a huge difference. The researchers write that “a firm can implement targeted promotions with customized timing using easily available descriptive statistics of households’ purchasing histories.”

Anyone who has ever walked out of a grocery store with a handful of coupons knows that marketers are watching what we buy. The evidence: the coupon for Mighty Dog that pops out of the register when you buy Little Cesar dog food. The marketers are hoping you'll be willing to try something new.

After all, consumers can be a fickle bunch. In fact, Sungho Park has found that they can be disloyal to entire food categories, not just brands. An assistant professor of marketing at the W. P. Carey School of Business, Park first noticed cyclical buying patterns close to home: his own home, that is.

“In South Korea, we didn’t have pretzel snacks with different flavors. When I first came to the United States for my Ph.D. study, I was obsessed with those snacks. Each time I went to the grocery store, I purchased them,” he recalls. “Then, all of a sudden, I found I wasn’t purchasing pretzel snacks anymore.”

Park says such episodes of food infatuation are common in many food categories, and it made him wonder: How common are these purchasing cycles, and how do they impact the effectiveness of sales promotions?

Using data from the SymphonyIRI Group, a market research organization formerly known as Information Resources Inc., Park decided to explore purchasing cycles more closely, and he teamed up with Sachin Gupta from Cornell University to do it. Their findings are in the May-June issue of Marketing Science, a journal published by INFORMS (Institute for Operations Research and the Management Sciences).

The scholars found that timing promotions based on buying cycles makes a huge difference. If markets get customers when they’re in high-buying mode, the product is far more likely to fly off the shelf. In fact, the researchers write that “a firm can implement targeted promotions with customized timing using easily available descriptive statistics of households’ purchasing histories.”

Peek perfection

Those SymphonyIRI data that Park examined provide a comprehensive peek into what participating consumers buy. The market research company recruits households to participate in its panel, and these people receive compensation in exchange for scanning the UPC codes on every item they bring home from the store. SymphonyIRI double checks home-scanned data against store-level data to verify the information collected. Then, household purchases are correlated with demographic information to give marketers detailed insight into the customers for their consumer packaged goods.

Such insights fuel classic target marketing, in which customers are segmented according to income, gender, family size and other demographic characteristics collected by companies like SymphonyIRI. Those data might tell you who usually buys your product, but they don’t account for the fact that, sometimes, folks switch from one thing to another, as happened with Dr. Park and his waning interest in pretzel snacks.

To probe purchasing cycles, Park and Gupta looked at SymphonyIRI histories of yogurt buyers in Eau Claire, Wisconsin during a two-year period that spanned 2003 and 2004. The team chose yogurt because it’s a highly perishable item, so consumers can’t buy large quantities at once and stockpile it for long periods of time. Among yogurt buyers of these three food categories, 73 percent had one or more 10-plus week periods where they passed on yogurt. The researchers chose to define a no-consumption period as 10 or more weeks without a purchase because they found that yogurt averages a 28-day shelf life in the fridge and, while some brave souls might eat it a week or two after the use-by date, few would eat it six weeks after its freshness expires.

The team also examined data for milk and frankfurter buys. In both categories, over two-thirds of buying households display at least one episode of several consecutive weeks of no purchases (6 or more weeks in milk and 18 or more weeks in frankfurters). This episode is long enough that in these perishable categories we can be reasonably certain there is no consumption as well.

The time is right

Of course, there also were times when people bought more or less of the watched food categories. The researchers used statistical models to track this and saw that 38 percent of consumers who regularly purchased yogurt went through high versus low buying cycles. That cyclicality, they reasoned, might impact the power of targeted promotions.

To test this theory, the scholars created a model to evaluate how timing played into the marketing mix. “You can waste money promoting your product to those cyclical consumers at the wrong time,” Park says. “Sometimes they’re not interested in yogurt because they are obsessed with apples or carrot juice. In my case, when I am addicted to potato chips, promotions for the pretzel category are not all that attractive.”

Using their statistical model to create a simulated promotional campaign, Park and Gupta tested how a 30 percent price cut might work if the promotion was timed to hit each household at a high point in its buying cycle. The simulation showed that offering households discounts in the midst of their high buying cycle delivered an 89 percent increase in promotional impact compared to the same offer sent at random points in the buying cycle.

But, how would marketers know where people are in their buying cycle? According to Park, this can be easily determined. He and Gupta created a predictor model that delivered a correct classification of where the household was in its buying cycle 63.5 percent of the time, or nearly double the rate that random guessing might deliver. Next, the team used these predictions and applied the simulation to see if the predicted high-buying households also would deliver more sales in response to the price cut. They did. There was a 46 percent improvement over randomly timed promotions.

Deeper look

According to researchers at Catalina Marketing, the industry-leading provider of those in-store coupon machines that deliver savings along with grocery receipts, an average shopper buys less than 2 percent of the UPC codes in the grocery store over the course of a year. No wonder marketers are so focused on finding and winning new buyers. New data acquisition tools and modeling approaches will likely help them out.

For instance, a recent New York Times article spotlighted analytical successes at Target, where statistical modeling helped marketers identify and woo women who are pregnant. Correlating baby-registry participation with purchases is just one way companies gain insight. The Times article quoted Target statistician Andrew Pole saying, “If you use a credit card or a coupon, or fill out a survey, or mail in a refund, or call the customer help line, or open an email we’ve sent you or visit our Web site, we’ll record it and link it to your Guest ID. We want to know everything we can.”

To that end, Park says some market researchers have started imbedding radio frequency identification (RFID) chips in loyalty cards and shopping carts. “They can see how you move through the store. If you’re a planned shopper, cross selling might be difficult in your case, but if you spend a long time in one place comparing different products, that may be a different story,” Park notes.

Who knows? Maybe lingering in the baby food aisle will someday tell Safeway you’re expecting long before the blessed event occurs. You’ll know that’s what’s going on when the coupon machine next to the cash register spits out discounts on pacifiers and diapers three months before a mom is even due.

Bottom line

• Even those who love a certain food category – such as yogurt or frankfurters – often have periods of no consumption, or they cycle between low and high purchasing patterns.
• In a recent study, 38 percent of yogurt-buying households cycled between high and low purchase periods.
• Marketers can easily track such patterns, and they should. Where cyclical buyers are in their purchasing pattern will impact the effectiveness of sales promotions.
• In a simulated sales promotion, promoting yogurt to consumers in the high point of their purchasing cycle lead to an 89 percent increase in the impact of the promotion.
• The researchers identify timing as a new dimension to targeted marketing decisions. According to the article, “a firm can implement targeted promotions with customized timing using easily available descriptive statistics of households’ purchasing histories.”