Since 1976, SAS has developed analytics software to extract intelligence from spectacularly large datasets.
Much of its dealings occur in financial services, where decisions around fraud, compliance and risk require trillions of computations. But the comparatively recent proliferation of online channels has created a need for rich analytics among marketers. And as the marketing and advertising worlds collide, that need is only growing more acute.
In February 2012, SAS acquired publisher-side ad server aiMatch, which it developed into its SAS Intelligent Advertising product. Wilson Raj, the company’s global customer intelligence director, spoke with AdExchanger about the coming conflation of advertising and marketing technologies, the data challenges companies will face bringing these two disciplines together and the future for SAS in advertising tech.
AdExchanger: How do advertisers vs. marketers handle data management?
WILSON RAJ: As it exists, data management is defined differently by these two camps. When you talk about data management from an advertising perspective, it’s typically more channel-specific. You have data vendors focusing on the TV audience, like Nielsen or Kantor, or social audience data like Bazaarvoice, or social, credit and financial data that Experian provides. But they’re not merged together. So that’s the advertising view of the world, from a data perspective currently. Even within that world, there’s fragmentation among all those that I mentioned.
We switch over to marketing data management, where the No. 1 source is CRM. This is all the historical data that pertains to sales interactions, venue interactions, Web behaviors.
Then when you merge [marketing and advertising], that’s when you have that holistic platform.
What’s the central challenge in that merging?
How does a customer go from anonymous to someone who is known? You have cookies and sessions that both advertisers and marketers track. You have data from browsing behavior, level of interaction, level of involvement. You have product recommendations for linking.
You start building this profile and moving forward it’s more behavioral. As you move further into the value chain, that’s when you get customer value. You get more predictive insight.
When moving from anonymous to a named account, the advertising side has more third-party data, things they’ve aggregated through credit institutions, specialty data, TV viewership, ratings, etc. What’s missing is the first-party data that typically marketers have. That’s the ID, especially if it’s a current customer. The question is how do you link those two together. One of the most difficult things to do is moving anonymized, aggregated data to more specific, individualized names. To figure out who’s interacting with you while they’re doing it, and understanding what their preferences are.
How do you link those two together without creeping people out?
We did a survey on that. In terms of privacy, 71% of 1,200 respondents in the US said that’s a problem for them, but 61% do want personalized experiences at the same time.
From a SAS perspective, we put protocols and securities in place to figure out how the data is used, in accordance with privacy and preferences of the customer. From an implementation perspective, there has to be an agreement or a process where if you’re working with advertisers, you need understanding of how you can extend security and privacy principles that are typically inherent in marketing, and extend that to the advertising side, and vice versa.
If you don’t have that overarching view of that entire dataset, then you will run into unwelcome kinds of interactions.
Who is responsible for overseeing this? The advertising agency? The brand marketers?
From what we’re observing, all of the vendors and all of the parties involved are coming to some sort of ownership of this. From an advertising perspective, the only way to grow now is to add more customer insight, add more accuracy in terms of attribution and add better measurements of performance. So they have an onus to own some of those processes.
The brand side is looking to additional data sources to enrich their first-party data, so they’re moving upstream.
From a vendor perspective, their motivation is to be a value-add, to be a trusted partner to the brand. They don’t want to become another specific channel for data that adds only limited value.
But the onus seems to be more on the marketer, on the CMO, representing the marketing organization as the steward for the customer experience. They’re more tasked to make these things works, but there’s still a shared responsibility with the other folks involved.
There’s a desire to move data-management platforms (DMPs), typically an advertising play, into the marketing stack.
That’s something that we definitely saw in February 2012 when we first talked to AdExchanger about [the acquisition of] aiMatch, an ad-serving platform that did a lot of things beyond single channel. That’s one of those opportunities we saw, not just from a technology perspective to bridge that gap, but we also saw the publishers wanting to add more value to their clients, the marketers.
When you acquired aiMatch, it was a sell-side ad server. Now it’s moving to the buy side of the house. Was that always SAS’ intention?
That was the intention. We had to start with the sell side because that’s what aiMatch was doing initially. The capabilities were directed toward real-time bidding, better leverage of exchanges, better business processes and so on. But we’ve always had an eye toward enriching those capabilities and connecting them to the customer intelligence solutions we already have. To get further penetration with the publishers you have to show the value.
One of the barriers we had to overcome was we had to help advertisers become more digitally and analytically savvy. On the campaign side, data-driven marketing had been – I won’t say operating at a high level – but it wasn’t a new concept to those guys.
It’s surprising to hear that advertisers aren’t as analytically savvy.
Yeah, I know. When you look at the sort of reporting that was done, typically, there’s a lot on aggregate information. You have impressions and views and click-throughs. But when it comes to deeper insight, it’s more descriptive. You know people saw it, people responded to it, but that’s at the aggregate level. When you get into why ads behave the way they do, what they’ll be doing in the future, any predictive kinds of things, those capabilities are not prevalent in the advertising side.
We see more of that on the marketing side, where people are getting into recommendation engines and predictive analytics, and there’s room to grow for sure. But data-driven marketers are already moving into that realm: next best offer, next best action. Those capabilities are not prevalent on the advertising side.
How many customers do you have using SAS Intelligent Advertising? How many are still in beta?
[As of this writing, SAS is checking to see if this information can be made publicly available.]
Are most of your current Intelligent Advertising customers new or were they inherited from the acquisition?
We’ve had quite a number of customers that we’ve acquired after the acquisition. Some of them are carrying on existing aiMatch customers, but we’ve added on a bunch of new companies, telco companies in Europe.
Do these companies use Intelligent Advertising as a standalone or have they integrated it with marketing tech?
They’re definitely seeing the connections between ad tech and marketing tech. [Some clients] using SAS for campaign management and marketing technologies are now tapping into components of AI to further enhance that personalization and relevance into digital channels.
Likewise, we had a customer in Europe where they’re using Intelligent Advertising components to do very innovative mobile marketing with some dual location capabilities and combining that with their existing campaign management solutions from SAS.
Can you elaborate?
I can’t name the account yet. It’s still a proof of concept we just completed recently and now the idea is to build it out and monetize it.
We had an internal case study in Europe for an online [clothing store] and an advertising agency. Those were the two entities we’re talking about here. In this case, they were able to use Intelligent Advertising, particularly the mobile analytics piece. The agency built an online banner ad for mobile devices and we added dual location capabilities and put together a mobile campaign for this online retailer. The customer fed in as much CRM data as they could: customer insight preferences and targets in terms of what inventory they wanted to sell. With that as a premise, an application was built by the advertising agency and they started targeting a few customers.
What happened?
In real time, the agency was able to get new information in terms of what the audience – in this case it was professional women in a European city – what kinds of boots they were clicking on. From an online retailer perspective, they were working with various POS retailers where they were selling their merchandise through. They got intelligence using geolocation where the prospects were moving. They could have timed offers as when the person was close to one of those stores, where they got an alert and a couple of offers on the phone. Based on click-throughs, the agency adjusted the creative, the offer type and even some terms of the discount, being able to figure out which were the best offers resonating to this particular category of women in the city.
This was sent back to the customer – the marketer in this case – where they were able to review it and adjust the campaign,the product they wanted to feature and the website.
So you’ve got an advertising campaign that’s merging and providing intelligence to the customer campaign and vice versa.
How long did it take to build this out?
This was a customized campaign. It took on both ends with help from SAS about a couple of weeks, almost a month, to get the infrastructure in place. Once we got it up and running, it was about optimizing and improving not just the process but the data capture and the analytics behind it.
Was your deal with the retailer or the advertising agency?
It started with the retail customer, but they had very strong ties with the advertiser so in this case, SAS felt from an ecosystem perspective and from a customer experience perspective, they needed to talk to the advertising agency. With that agreement, the agency saw value and it drove better ad targeting and ad efficiency, and certainly a deeper relationship with their customer the online retailer and vice versa. And the retailer took advantage of the rich third-party data the advertiser had and their know-how on mobile advertising and so on. What moved them together was the analytics and the orchestration process for both sides.
You mentioned this is a proof of concept you’re looking to monetize. What are the goals there? Getting customers to make more purchases?
The proof of concept had a sell-through component, but now we need to scale it to more full-production mode and expand it beyond just geolocation. We want to take it to multiple cities and several countries, and take it to other retailers and agencies that might be interested.
When it comes to data, what aren’t advertisers thinking about that they should be thinking about?
First, expand their perception of data and go beyond what they’re already collecting, to partnering more with their customers – in this case the marketers – to aggregate, match, cleanse, and get to the point where you can have a nice set from anonymous to known customer data.
Second, be more analytically driven. That’s something we see in the news quite a bit where the huge mega-agencies and the smaller scrappier ones talk about analytics. But they have to look at it not just from a media optimization perspective but also to use things like predictive analytics and predictive principles to anticipate audience behavior based on past data, present data and to forecast future performance.
Third, embrace technology a little more. Not that they currently aren’t, but when you look at these ecosystems, there is no one solution or one vendor that can provide everything. It really needs to be a partnership like that example in Europe, where you have the online dealer meshing with the advertising side, and having a common data platform and a common analytics platform to execute on those things.