AdExchanger: To what extent does the influx of data coming from a bunch of different channels affect your messaging?
JEFF MALING: In my opinion, there’s a lot of data and not a lot of insight. Most of our clients are not leveraging the data they have to push customer conversions. There’s still a lot of talk but not a lot of doing. When we’ve run pilots around improving the experience from when you search all the way through conversion and improve the relevance of content in that whole change, the results are stratospheric.
Give me an example.
JM: Say I’m looking for a car and I want to go to Napa. I’ll type “Napa convertible rental" into Google. Most experiences today, that’s the last time anyone knows what I really want. Maybe there’s a banner ad or search term I click on, but now you’ve lost my context. So when I get to Hertz, there’s no clue what I‘m trying to accomplish. We’ve found if you can keep contextual relevance through that entire cycle when I get to Avis, I understand you want a car rental in Napa. I know maybe you’re not as price-sensitive, because you want a convertible. I know that you’re going for a weekend, so it’s leisure travel. I know a lot about you based on what you’ve typed in and the journey you’ve been on.
I can take that data and do multivariate testing, and I can run different creative for that particular persona. Car rental convertible. Napa convertible. That’s a microsegment, but to pick something bigger: one-way car rental. Car in the shop. Being able to use that data in the process, when done right, changes the conversion metrics. We’ve done a lot of dynamic creative for GM as well and the numbers are very positive.
To accomplish that, do you work with attribution vendors or work with an in-house solution?
JM: All of the above. We have our own attribution capabilities. We have some tools around dynamic landing pages to improve the relevance and experience as they go through it. We also work with outside vendors and our sister companies, like [Dentsu Aegis Network-owned media agency] Carat, who can deliver a lot of information.
ABEL REIS: We have a lot of clients inside the Brazilian portfolio and we set an agreement recently with them in that we’re capturing cookies for each visit, each consumer. We’re creating an Isobar cookie pool that’s helping us deliver more customizable experiences in the future and basically using a first-party DMP for programmatic movie, in partnership with [Dentsu Aegis Network-owned trading desk] Amnet.
I didn’t know you had your own cookie pool in Brazil.
AR: As an experiment.
Do full-service digital agencies typically have their own cookie pool?
AR: We are living in exciting times. The frontier is not so clear. The lines are a little bit blurred. We have a lot of challenges and, in Brazil specifically, the clients ask us about how to handle this massive amount of data that I’m capturing every month from clients visiting the site.
And when you have a situation like that, sometimes the lines get a little bit blurred. While data is typically something that media publishers have in their hands, if we manage the digital presence of our clients and have thousands and thousands and millions of visitors, we have to put those assets to work.
Would you consider developing an Isobar US cookie pool?
JM: Tools help clients better understand the data they’re harnessing with their customers, and that’s important. Whether we do that as a body within the client or one outside of the clients is dependent.
We developed Radar, a media metrics product. We use that with our media sister companies Carat and Vizeum, and we’ve rolled it out to Macy’s, Home Depot, Smuckers, GM. And we’ve found that competition aside, an interesting dialog takes place among those clients. Does that eventually transfer into a shared offering? We aren’t quite there yet, but it might.
How different is Brazil compared to the US?
AR: Brazil is a different market than the others. Usually agencies in Brazil are responsible for creativity and media at the same time. Which is quite dangerous in the digital landscape. In Brazil, Isobar is the biggest digital media buyer in the country. That means we buy more than all of the conventional or traditional agencies that are running in Brazil. Media is a key part of all business in Brazil.
Why is mixing media and creativity dangerous?
AR: We have to adapt our operations to the local regulations and conditions. In Brazil, (independent) media agencies are an entity that cannot work because there are a lot of legal restrictions. That’s the reason Isobar in Brazil runs two tracks: One is media and the other is creativity and technology.
But I have to say it’s not bad to have media, creativity and technology in the same box. Because we can explore the opportunities in the partnership with media publishers and technology providers to empower the final deliverable.
AR: We developed for [Brazilian subscription television service provider] SKY an approach where you use your cell phone to program your box in your house to record a program that you want to see later. And we are using Twitter to do this. It’s interesting because Twitter is in the social space, but we use Twitter as infrastructure to control a satellite. When you tweet “please record the next Simpsons episode,” you are controlling the satellite, your box and are broadcasting to your friends that you will see Simpsons tonight.
JM: So if someone says on Twitter, “Just watch this episode of the Simpsons, it’s crazy funny,” you can hit #record. And that then says, “Abel is recording this episode of the Simpsons.” That dialog is happening on Twitter because the broadcasters are seeing all this traffic.
AR: And at same time, we set a massive agreement with Twitter. We are looking at Twitter as a technical infrastructure, as a social space and as a media publisher. At the same time doing a great business for us and for SKY.
Where else is Isobar a big media player?
JM: In China we’re a big media buyer for digital. The regulations are similar, as they are with a few other Asian countries. Europe and the US, because of the regulation structure, it’s a specialist space. Isobar US is not a media buyer. We’ll partner with Carat or Vizeum or the client’s media specialist.
Why would there be regulations to tie media and creative together?
AR: Well. Next question, please?
JM: We call it Planet Brazil.
AR: It has to do with our history. In Latin America, in Brazil specifically, you have a few media companies having the control of several channels. They have interests. So if you look for the reasons things are as they are, it has to do with initiatives to protect the market from outsiders. I wouldn’t say it’s wrong, but it’s a fact of life. It’s not the most important aspect of our conversation or our business. We have to adapt to each market, trying to deliver the same mission, same vision, same proposal, considering the local peculiarities.
JM: It’s not that different than China having different social media platforms. It’s just a fact of life you have to deal with Weibo and WeChat and all the other ones that are there.
Sometimes there’s tension when combining creative and data-driven concepts. Why?
JM: I don’t necessarily get that. Why wouldn’t you use all the data at your disposal to inform the creative process? For a car, you might gather a lot of data and realize no one likes a feature. The issue could be the feature sucks, or the issue could be consumers don’t understand it.
AR: It helps the business guys to know a feature doesn’t make sense because people aren’t looking for it.
JM: Or at this part of the sales process, you’re focused on the eco-friendliness of this car and it’s not working. The thing that data gives you is much more crispness on the problem that you’re trying to solve. All the creatives we work with like a real crisp description of the problem. And once they get that problem, they’re real good at coming up with solutions for it.
How do you ensure causation?
JM: You can upfront test. And we like to use multivariate testing on the platforms as well. On Avis, where we’ve done our most advanced testing, they wanted to go out with the prepaid function. The issue with car rentals is they don’t make you pay for it ahead of time. Wouldn’t it be great if they can get you to prepay? They’d get more money up front, and it’s more likely customers will show up, which is a boon for the inventory.
So what offer would they have to give to get someone to prepay? We used multivariate testing to test the elasticity of demand. Surprisingly the actual number was about three times less than what our client would have guessed.
They were way overestimating the impact the function of prepaying would have on their rental. People were willing to do it for a pretty nominal discount, which dropped a lot of money to the bottom line.
Looking at correlation-causation, you have to ask if you really know or just think you know. And if you just think you know, you need to do testing. You can do market testing or multivariate testing, but you need to test.
Edit 6/24: Some of the specific figures around Avis and Royal Caribbean were meant to be illustrative and weren't exact. For reasons of accuracy, these have been edited.