Practical Ecommerce

Beware retargeting

Advertising is a pain in the butt. It’s expensive. It requires constant maintenance, and it’s not always measurable. In our industry, though, we tend to think that we can determine the ROI of everything. Truth be told, that has become a bit of a fallacy due to the increased number of advertising vehicles.

First, let’s examine the most popular advertising opportunities available to most of us who sell online.

Pay-per-click is the most measurable and arguably the most effective. It’s also the most static. Unless more people search your terms, you cannot expand this area of your business. You may be able to tweak your cost per conversion, but you can’t simply double your PPC spend and expect to double your profit.

Of course, most of us are okay with this. PPC is the platform upon which we build. It brings us new customers that hopefully we can convert into repeat buyers. It provides a steady flow of business and we’re generally happy with it.

Display advertising sucks. I hate display advertising. I’ve thrown so much money down the drain on display advertising. But every few months I always think I can “crack” it. You can’t. Display advertising will never work for direct response. It is a branding medium only. If you’re selling memberships, it’s great. If you’re selling products, stay away. Display ads are too expensive (more on this below) to be cost effective unless the lifetime value of your customer is quite high.

And then there’s retargeting, the supposed savior of display advertising. It promises so much and backs it up with fishy math. Retargeting is the process by which you show display ads on other online properties after a visitor to your site has performed some action. If they’ve abandoned your cart, you can show them ads for your store when they visit Facebook, YouTube, or any other site that participates in your retargeting network.

Retargeting is a huge reason why display advertising is so darned expensive. Most of the display inventory is being bid on by retargeters who are willing to pay a lot more to get in front of their customer than you are.

Retargeting sounds great, but the problem is measurement. Therefore, as long as you make your retargeting decisions with full understanding of the fishy math, you’ll be okay.

When it comes to cost per conversion, how do you attribute the sales? Every advertising source wants to take account for a sale if it had a hand in securing it.

To illustrate the problem, let’s examine the following scenario. A customer searches for “beach ball” on Google. She clicks on your PPC ad. She explores your site, adds a product to their cart, but then leaves before completing the purchase. She then hops on over to Facebook. She is shown one of your retargeting ads. She doesn’t click on it, but it was displayed to her. The next day, she receives an email from your shopping cart abandonment service. She clicks on the email and completes the purchase.

In that scenario, who gets credit for the sale? In short, everyone. And that’s the problem. The PPC engine takes credit, your retargeting engine takes credit, and your shopping cart abandonment engine takes credit. Yikes.

Let’s say that your cost per conversion on your PPC engine is $7, your retargeting engine is $5, and your shopping cart abandonment engine is also $5. Did you just pay $17 for that order? Nobody truly knows, but it’s possible. More likely, though, it’s somewhere in between $5 and $17. However, if you only view the reports from each individual service, you’ll be misguided.

There are a lot of sophisticated algorithms and software services that supposedly help you understand the complex problem of attribution. They assign weights to each marketing engine, but this is all designed for large corporations. For us small guys, we don’t have the money or the time to invest in these tools. Plus, they’re a waste of time.

Use some common sense to understand attribution. You know how much you spent on advertising last month and you know how many orders you got. That is the maximum cost you paid to get each order. Don’t over complicate it.

Of course, this back-of-the-napkin math doesn’t help you in determining which marketing engine you should pump more dollars. For this, well, you have to make some assumptions.

I assume that clicks equal intent. Therefore, I attribute a lot more weight to an advertisement that has been clicked rather than simply viewed.

When it comes to shopping cart abandonment emails, I only go by clicks. Because everyone who abandons his cart gets an email, I don’t attribute every order whose cart was previously abandoned to be a result of having seen the email. I only attribute a click on the email to a conversion.

Retargeting is a bit trickier when it comes to assigning weight. Again, I believe that clicks are far more valuable, but there is enough data to suggest that “view through conversions” are a legitimate metric. VTC are conversions attributed when a customer is shown one of your retargeting ads and later makes a purchase.

However, would customers have made the purchase anyway without seeing the ad? Sure, some would. But some wouldn’t have. This difference is known as “lift.” The only way to determine the lift is to perform an A/B lift test with your retargeting vendor using your ad and, perhaps, a public service type ad. In my experience, the lift is generally positive, but only through tests can you determine it definitely.

Illogically, popular vendors like AdRoll require a $5,000 / month spend in order to perform a lift test. This should give all of you pause when viewing AdRoll’s inflated ROI metrics.

The bottom line is that the more vehicles you use, the less you can rely on their cost per conversion data. Determine what you value the most (clicks versus impressions) and spend accordingly. But don’t take the cost per conversion values at face value.


Get the Practical Ecommerce RSS feed

Comments ( 6 )

  1. david mcbee March 6, 2014 Reply

    Full disclosure, I work in the display retargeting space. That said, your title grabbed my attention and I came to your article ready for battle. Instead, I found this wonderful quote: “Use some common sense to understand attribution. You know how much you spent on advertising last month and you know how many orders you got. That is the maximum cost you paid to get each order. Don’t over complicate it.”

    I believe that a lot of time is spent trying to determine what advertising methods work, but that it can be impossible to correctly attribute the success of a campaign by measuring clicks.

    quick story, when I sold yellow pages 100 years ago, we would often put tracking numbers in the ads. When the advertiser also ran newspaper, TV or radio (way less engagement – read click – than yellow pages), we found that our call volumes increased significantly. And yellow pages got all the credit when this happened. This isn’t unlike a display ad influencing a google search that ends up being a PPC click.

    So I agree. Use common sense. Measure the lift of traffic to your site, phone calls, and most importantly – sales. If the blend of strategies is working, you’re probably doing something right – even if you can’t drive it down to an exact metric.

    • Jamie Salvatori March 6, 2014 Reply

      David, thank you for the comment. I’m glad my headline worked as planned!

      I think retargeting works. But I don’t think it works as well as every retargeting platform’s dashboard indicates.

      Of course, they’re not to blame. If their ad has been clicked and a conversion takes place afterward, sure, they can take some of the credit — how much is up for debate.

      The VTC’s are a bit trickier (unless the platform gives you an A/B lift test). Even that being said, you must assume that you probably paid to get the customer into the retargeting funnel in the first place. So, yeah, it’s tricky. All advertising works, but determining which is working is probably impossible and that’s why the big boys do a bit of everything.

  2. Charles Nicholls March 11, 2014 Reply

    Hi Jamie

    I was also hooked by your headline. And in full disclosure, my company runs remarketing using email and ads for more than 4,000 brands.

    I’m really glad that you chose to throw the spotlight on this thorny area.

    You raise a lot of good points about attribution models being flawed – and I couldn’t agree more. The old saying of ‘Half the money I spend on advertising is wasted, the trouble is knowing which half” couldn’t be more true today. With digital advertising of course the expectation is that it will be measurable. But as you point out so eloquently, in reality most of these measurements are overlapping over-estimates.

    Our approach to attribution is different – we measure everything as lift over a control group using a continuous A/B test. This approach is an accurate way to measure the effects of remarketing, as opposed to estimating it using a click + attribution. The control group measures all of the other influences on your customer, such as promotions, other marketing campaigns, competitor activity etc and allows these to be subtracted from the campaign measurement in order to measure incrementality.

    The other big benefit from A/B lift tests is that you can see when campaigns are not working – and this is critical in understanding the true drivers of performance. For example, do you make more money showing ads and sending emails to cart abandoners? Or are you better off sending only emails where you’ve captured an email address, and saving the ad budget for the unidentified visitors? Unless you test it you won’t know.

    Great article. Thanks for sharing.

  3. Andy March 11, 2014 Reply

    Amen brother! Perhaps a better name for the article would be beware attribution. There is so much snake oil and BS being pushed in the big data digital marketing world right now it is almost criminal. it makes my job very difficult when decision makers are bombarded with miracle solutions every day. Nothing exciting about pragmatic testing and common sense. This article is right on point in describing my frustrations. Thanks you. It is nice to know I am not the only one.

  4. Ofer Tal March 18, 2014 Reply

    Excellent summary of the issues I recently witness. It is sometimes funny (or painful) for online retailers to see the overlapping credit their advertising agencies are taking on the same orders. The ‘who gets the credit’ doesn’t match the ‘why the customer finalized the purchase’.
    Measure the clicks is the simple and clever take, can’t agree more.

  5. Frank Rumbauskas April 30, 2014 Reply

    Excellent article, and I’m with you 100% on this.

    I’ve tried multiple retargeting platforms. I’ve also invested tons and tons – and more tons – into email marketing.

    In 5+ years of running retargeting ads I still have not been able to attribute one lone single solitary conversion to the ad, rather than my very finely tuned email follow-up strategy.

    As a sales author/trainer I always think of retargeting as the salesperson who badgers you relentlessly after you’ve said no. Because one in a hundred of those prospects will eventually buy, salespeople buy into the myth that “persistence pays” but in reality it just takes time away from better prospects.

    Likewise, my opinion on retargeting is that you’re chasing prospective customers who are better followed-up with by more proven mediums like email, or even good old fashioned direct mail or a phone call.

    Frank Bettger’s “How I Raised Myself from Failure to Success in Selling” describes how 95% of his sales took place within the first two appointments, yet he spent most of his time going to appointments 3 and beyond with prospects. He tripled his income when he decided to abandon people who didn’t buy after the 2nd appointment.

    People who are drinking a bit too much of the retargeting kool-aid would be well served to take that lesson to heart!

Email Newsletter Signup

Sign up to receive EcommerceNotes,
our acclaimed email newsletter.

And receive a free copy of our ebook
50 Great Ecommerce Ideas