10 Minutes with Kirthi Kalyanam

Hal Conick
Marketing Insights
Current average rating    
Key Takeaways

What? Kirthi Kalyanam ​and his researchers​ had an ah-ha moment: a high frequency of ads is more effective in remarketing campaigns.

So what? "That’s a pretty powerful realization for us. In the future, there are going to be tremendous opportunities to study the impact of marketing on existing customers or people you actually know something about versus people you know nothing about," Kalyanam says. 

Now what? While this is just one study, Kalyanam says he expects future studies to confirm these numbers and hopes best practices will change. 

​Nov. 17, 2016

One marketing professor believes his new study is an “ah-ha” moment for the question of frequency in remarketing


Remarketing has become a necessary tool in marketing, but even seasoned professionals aren’t sure which frequency works best. Will high frequency of ads every day annoy potential customers? Will low frequency go unnoticed?

Kirthi Kalyanam believes he and fellow researchers had an aha moment in a recent study.

The “The Timing & Frequency of Retargeting: Large Scale Field Experiments at BuildDirect.com,” which was presented at Omni.Digital in Chicago in September, studied daily impression frequency caps of zero, five, and fifteen for BuildDirect’s website. The study’s results were resoundingly in favor of high-frequency remarketing, which beat zero-ad- and five-ad-per-day frequency in effectiveness each of the four weeks studied.

Researchers found timing is also an important factor, as 85% of a remarketing campaign’s effect is felt within the first four days.

“Our informed hypothesis is this is due to competitive interferences,” Kalyanam says. “When people are shopping, they look at other people’s websites and go there. Because of that, everyone starts retargeting them. Your ability to reach them becomes weaker and weaker within four days. Because of that, if they didn’t reach you in week one, week two is not effective.”

Marketing News spoke with Kalyanam about remarketing’s evolution, its future and a possible change in how remarketing frequency is viewed by marketers.

Q: Can you explain your study, as well as why you wanted to look at the frequency of remarketing?

A: Remarketing started out very narrowly as something called retargeting. Retargeting became extraordinarily popular on the web when people observed you came to a website and after that, if you didn’t do something they wanted you to do, they would come back and advertise to you. That was the original kind of retargeting.

As you may recall, that conveyed all the display inventory into intent inventory. Before that, display was being shown to people based on demographics and there was no behavioral intent behind it. Now all of the sudden, we could connect display inventory to intent, an extraordinary innovation. …

We have evolved to … before I show you advertising, I’m going to show advertising to people who I know have been on the website, and I have some contact with, versus people I have not [had contact with]. That’s now spread to other areas. Of course, e-mail is an area where if you have my purchasing e-mail address, conceptually, you should be able to do different types of remarketing to me. It’s also becoming more popular in search advertising where platforms like Google allow you to do more and more remarketing. Facebook is doing some of the stuff too. This whole area has become quite interesting.

Q: So marketers can have a better grasp of who they’re marketing to now?

A: Correct. It’s ironically connected to the famous saying, “Half the money I spend on advertising is wasted; the trouble is I don't know which half.” Remember that famous quote from John Wanamaker? I think what he was saying was one half of advertising was wasted because I don’t know if it works, but now I think I’m advertising to people who are already my customers, but I’m still not sure. So [the remarketing issue] captures that extremely well-known saying in a new light.

That’s a pretty powerful realization for us. In the future, there are going to be tremendous opportunities to study the impact of marketing on existing customers or people you actually know something about versus people you know nothing about. This whole notion of remarketing is going to become broader in scope. Just because someone is a customer doesn’t mean you should stop marketing to them. It’s a very competitive landscape, and if you don’t continue to market to them, your competitors will start chipping away at you.

Q: Right, and return customers are your most valuable, from what I understand. You need them to keep coming back.

A: Yes. The frequency was something that both practitioners and academics became very interested in, because it was a well-known issue in advertising about how much frequency you need before you get people’s attention. Now, with internet-based advertising, you actually can look on the individual level at how some people get different frequencies than others and start understanding how frequency works. That drew our attention.

Q: Was it surprising to find in your study that high frequency paid off? That seems to run contrary to what many marketers think. I’m sure you’re familiar with the popular feeling of not wanting to annoy customers. 

A: We have done discovered something very unique that will also require additional work. It can be different in different situations. One issue is philosophically, how do you make sure you hit the right frequency? There’s a big concern out there that higher frequency will be annoying to your customers, but at least to the extent we’ve looked at it, we don’t see much evidence of that yet.

The third thing in our study is the timing, we looked at remarketing and frequency, but also timing. The timing issue was should I hit you right away or should I wait? And that’s a very powerful idea. You come to the site, don’t buy something and leave, should I retarget you right away or should I wait?

Q: I thought it was interesting to see that without remarketing in week one, week two’s effectiveness saw huge drop.

A: That is powerful. … Quite simply, it is a reminder that people need to forget, which means it works better later than earlier.

Q: Do you mean remarketing works better later in the day? If a customer went to a website early in the day, for example, would it pay off to not send a remarketing ad until evening?

A: Yes. In general, later in evenings, later in the week, later in another week. How long does it take for you to forget? If you’re going to commit very quickly, within hours, then we should reach you within hours. If you’re going to commit within days we should reach you in days. [The customers] need to forget, that’s all.

Q: Do you see this study changing how the industry works? Or perhaps starting a chain of other remarketing studies?

A: I did not expect the amount of excitement that this presentation caused. After I presented it, I got about 20 e-mails from people in the remarketing business who are doing programmatic ad sales [or] doing analytics. Once we had demonstrated a fairly nice, robust methodology to run these kinds of experiments and the extremely powerful insights you get out of it, people look at it like, “We should be doing this for our clients.”

Recommended For You:
Understanding the What and Why of Ad Retargeting Gain a Global Perspective at AMA Events   E-book: 2016 AMA Annual Conference Recap

Understanding the What and Why of Ad Retargeting

Gain a Global Perspective at AMA Events

E-book: 2016 AMA Annual Conference Recap


 Sign Up for Marketing News Weekly

Get the best marketing thought leadership delivered directly to your inbox!

Author Bio:

Hal Conick
Hal Conick is a staff writer for the AMA’s magazines and e-newsletters. He can be reached at hconick@ama.org or on Twitter at @HalConick.
Add A Comment :

Become a Member
Access our innovative members-only resources and tools to further your marketing practice.