Past browsing does not indicate future purchase

Jun 25, 2018
Doug Garnett

Through a special arrangement, presented here for discussion is a summary of a current article from Doug Garnett’s blog.

After years of articles about personalization efforts falling short, perhaps it’s time we realize it’s not the fault of retailers; it’s a faulty idea.

The assumptions about delivering “meaningful” or “personalized” ads based on past browsing (or purchase) ignore human realities.

People have tons of interests and we shift between them quickly. Even if I searched for a trash can last week, there are many reasons that ads about trash cans won’t be personal or meaningful this week. Something is only “personal” for an instant of time. As soon as the second-hand ticks to the next hash mark, it’s no longer personal.

The problem is made worse because online shopping is more about “buying” — not “shopping.” We go to the web to buy, and quite often do so very quickly. As a result, by the time algorithms see that I bought a trash can online and decide to send me ads about them, any shopping interest I have is for something other than trash cans. And it’s highly likely by then what I’m doing on the web isn’t shopping-related at all — my mind is no longer in that space.

Yes, perhaps there are enthusiast categories that I search (say, fly-fishing supplies). Except, I’m human. While I might want to look at fly-fishing stuff today on a coffee break, tonight I’ll need to buy a staple for the household and tomorrow I’ll order a new clamp for some work I do in the garage.

So, when your agency’s algorithms decide something makes an ad (or email or product recommendation) “relevant,” it’s highly likely they’re wrong. You run a very high risk of pissing me off and damaging your brand.

That said, online direct response marketers do find that digital retargeting is often a cost-effective way to generate online sales.

For every 1000 retargeted ads they feed, a few go to people who are interested and maybe one of those buys. These marketers don’t really care that their ads were irrelevant or meaningless to over 900 of the people who saw them — as long as the cost per order is right.

If we take care, there are great options for advertising online. Seeing them takes putting a critical eye on what we’re told are “common sense” assumptions about targeting.

DISCUSSION QUESTIONS: Do you agree that basing targeting on online browsing and purchase history is highly limited as a personalization method? Where do you see the value of browsing and online purchase data in marketing efforts?

Please practice The RetailWire Golden Rule when submitting your comments.
"Sure past purchase behavior is a great indicator of future buying habits but don't count browsing down and out either."
"Retargeting is much easier at the brand level than the product level. This avoids the “I already purchased this” reaction..."
"The Hebrew prophet Jeremiah wrote that the human heart will trick you every time, no one can know it. Algorithms or not, that still stands true."

Join the Discussion!

20 Comments on "Past browsing does not indicate future purchase"

Sort by:   newest | oldest | most voted
Lyle Bunn (Ph.D. Hon)

The concept has strength from a consumer experience perspective, but the cycle time and messaging algorithms need lots of refinement. Ads for products already purchased are useless, but cross-sell ads can be valuable. A barrage of ads based on browsing history becomes an annoyance quickly, as much as not seeing ads for products from recent searches. There is a lot of refinement yet to take place.

Dr. Stephen Needel

Perhaps because personalization is the wrong concept. Stealing here from Joel Rubinson – perhaps relevancy is the guiding concept. If I’m searching for a trash can, showing me a trash can ad is relevant. If I’ve just bought one, it’s now irrelevant. It’s only when purchasing is on a regular schedule (like dog food for my dogs, every 40 days) or there is a pattern — I bought four things, all with a Santa Fe Southwestern theme — that this attempt at personalization should work.

Ralph Jacobson

Browsing is only one piece of personalization. And it must be tempered with myriad other data points in order to paint an accurate picture of the shopper. I’ve shared this video before, however it is a great example of what happens in this situation.

Art Suriano

This article proves that technology is far from perfect and we have a long way to go. Unfortunately, too many retailers don’t care, and they run the risk of annoying customers. Online marketers need to do a better job on how they qualify a customer’s needs. For example, if I bought the item why continue to send me ads? That is the most frustrating, and if nothing more a simple “yes” or “no” question asking if I’m still interested would be excellent.

What I see as the more significant problem is that too many retailers believe that technology is the answer for all their needs and it is not. They give way too much freedom to the tech companies assuming that they know best when in fact too often they don’t.

Listen to your customers and follow what they want and then make sure your tech companies and in-house departments are complying with those needs. The customer comes first whether it’s in-store or online. Forget that rule and you’ll never be successful.

Tom Erskine
1 year 7 months ago

Yes it does.

Personalization is all about understanding context, and while the author is right that one data point — searching for trash cans — doesn’t provide enough context to predict future behavior, when you combine multiple data points from my browsing and search history you can get a pretty good picture.

If I search for trash cans, but also hoses, painting contractors and a new dishwasher then you might accurately predict I just bought a house. Or if I search for cribs and baby monitors you might want to advertise other baby gear.

Shep Hyken

When it comes to personalization, there is a big difference between browsing history and purchase history. No doubt there are patterns of a browser that can almost guarantee the consumer has a genuine interest in a product, but it’s purchasing history that will give the greatest indication about what the person is interested in buying.

Charles Dimov

There is definitely room for the next major breakthrough in marketing technology. Technology which identifies that a shopper has already purchased an item and suggests related products. Lyle is on to something in his comment.

Presuming you don’t violate privacy policies and legislation — online browsing and online purchase data can help profile customers (yes I know, eerie). Yet, this will help cater products to me that I have NOT already purchased. There are miles to go before we sleep. We are definitely not there yet!

Celeste C. Giampetro

Let’s put a finer point on “browsing.” If a consumer visited the home page and then bounced then no, personalization probably won’t do much to persuade them to buy. However, if that same consumer went deep into the site’s product pages, put a few items in the cart and then bounced, personalization has a strong change of converting them.

As others have pointed out, frequency of the ad message is clutch. Inundating that consumer with reminders might not work but a respectful one-time message to incentivize them to return to your site may do the trick. Sure past purchase behavior is a great indicator of future buying habits but don’t count browsing down and out either.

Brandon Rael

Driving personalized shopping experiences based purely on browsing history is a legacy approach that does not consider the many physical and digital touchpoints along the consumer shopping journey. We are bombarded by targeted ads on social media platforms, almost immediately after visiting a retailer or service provider’s website. Browsing may simply be for research or comparison shopping and does not necessarily lead to intent to buy.

Leveraging browsing data is a valuable part of capturing the customer journey, yet retailers have to also consider that the journey extends well beyond the internet. With the lines blurring between the physical, digital and increasingly social media worlds, tracking the customer journey and driving personalization is an increasingly complex task, yet so critical for retailers to understand.

Evan Snively

Retargeting is much easier at the brand level than the product level. This avoids the “I already purchased this” reaction and likely still showcases meaningful messaging (albeit less so than a specific targeted product) to the consumer. Brands that sell products that are more a choice of fashion than function will have better success with this — getting a banner ad from Uline after I bought their trash can probably won’t have the same impact as getting one from Patagonia after buying a better sweater.

Ian Percy

This is THE most sensible article on the whole “personalization” issue I’ve read. Thanks Doug! I’m glad to see that my thoughts are redundant to most of my colleagues. The Hebrew prophet Jeremiah wrote that the human heart will trick you every time, no one can know it. Algorithms or not, that still stands true.

“Browsing” is an online synonym for curiosity. This morning over coffee I browsed for a KIA Stinger out of nothing but curiosity because a friend mentioned it at dinner last night. Now I’m braced for the “personalized” deluge!

Doug Garnett

Thanks, Ian… And I love the Jeremiah observation. We are dealing with the human here – in all its fully irrational glory. I hate it when we marketers presume we can ever fully understand that.

Anne Howe

Personalization, if done well, assumes there’s a two-way conversation between two parties that is about a person saying “help me with X.” Marketers can then start to use data to predict and presume things that will be helpful to the person more often than not. Ideally, when humans are involved, it gets better! Too many marketers are trying to take the easy road, and browsing history doesn’t necessarily correlate well with shopping intent!

Adrian Weidmann

The value of understanding online browsing and purchase history isn’t so much in individual products but rather in analyzing the trends and interests. Understanding that a particular person is interested in fly-fishing unlocks a broad spectrum of related attributes — outdoors, locations, books. If someone buys a fly rod, and all you target is fishing rods, you’ve already missed.

As the great one Wayne Gretzky stated — “I skate to where the puck is going to be, not to where it has been.” Marketers need to use AI and recommendation engines to develop a broad spectrum of insights and market to where shoppers are going, not to where they’ve been.

Lee Kent

There are many outliers when it comes to browsing and searching. Think about how many times you are searching for a gift for someone else. When is the next time you will be searching for that person and how is it relevant to you? This is why personalization in this way is difficult. I opt for AI over targeted ads. Much of online shopping is in the now. AI can address the now. What are you looking for now and what are the obstacles in getting to the purchase? Targeted ads can be very hit or miss — mostly miss for me since i rarely shop for myself. Now if you can detect that a consumer is a regular buyer of certain goods and at a certain frequency, targeted ads can work. However, as Tom said, “that involves multiple data points. Understanding context.” And that’s my 2 cents too.

Ricardo Belmar

This is where technologies that can better incorporate online and physical experiences have room to shine. Browsing history is one thing, and purchase history is another. However, how many times have BrainTrust members pointed out how crazy some recommendations on Amazon have been in their experience? This really only works for product categories that lend themselves to personal emotional value. I doubt anyone gets a strong emotional response from buying a trash can (or any related products, like trash bags). So yes, it really does nothing to improve the customer experience to target someone with trash can or trash bag ads just from browsing or even purchase history. Product categories like apparel (but not necessarily all), electronics, beauty, and wellness, are just some examples where this level of personalization and retargeting may be more effective.

Craig Sundstrom

Perfect? No. But “highly limited”? Even improving the strike rate from, say, 0.1% to 1% could be important, but the number obviously would still be very small, so I don’t think targeting should be written off. What alternatives are there? We should just adjust our expectations accordingly.

Patricia Vekich Waldron

Relying only history (consumers’ browsing or past sales) for personalization and planning limits you to a point in time and builds in the same errors.

Kenneth Leung

Browsing provides indication of interest, but you need context in terms of the product and the customer. Someone browsing and buying a toilet seat is not an indicator of immediate future purchase most of the time, browsing and buying a laundry detergent may indicate so in the future depending on how big the package is and how long the estimated consumption is. Having just browsing and purchase history for personalization is not enough for a good customer experience.

Susan Viamari

The time consumers spend online provides valuable opportunities to engage and influence, even for off-line purchases. But consumers don’t engage in ads that are irrelevant. Sure, browsing is a good indicator of purchase intent, but past purchase behavior is even better. The key, of course, is understanding the buy cycle. If you purchased a garbage can last night, you are very unlikely to purchase one today (or tomorrow). Marketers need to understand the purchase cycle for their products and use contextually relevant ads–that means timing, along with messaging and a value proposition, to really be impactful. Personalized, purchase-based targeting—ensuring that the right message reaches the right consumer at the right time and place—has shown to increase ROAS by 20 percent and enhance lift threefold.

"Sure past purchase behavior is a great indicator of future buying habits but don't count browsing down and out either."
"Retargeting is much easier at the brand level than the product level. This avoids the “I already purchased this” reaction..."
"The Hebrew prophet Jeremiah wrote that the human heart will trick you every time, no one can know it. Algorithms or not, that still stands true."

Take Our Instant Poll

Do you agree that targeting based on online browsing and purchases is highly limited as a personalization method?

View Results

Loading ... Loading ...