Among the first wave of digital agencies launched in the early to mid '90s, Razorfish may be the strongest survivor. Through several acquisitions that ultimately placed it in the stewardship of Microsoft, then Publicis Groupe, the agency retained a distinct identity while continuing to innovate and expand globally. Last week it grew its China headcount to 330 through the acquisition of Longtuo, an e-commerce enabler in the region.
Part of that identity is builder. Razorfish over the years has developed a sprawling family of in-house technologies, geared to a wide range of client needs. These tools cover ad targeting, content publishing, data visualization, and APIs/connectors. The products are the farthest thing from what you could call a "stack," yet Razorfish has decided to unify them under a single unit, called Fluent.
From a display ad standpoint, it almost feels like the second coming of ATOM Systems, the innovative but ill-fated trading desk which was folded into Vivaki's Audience on Demand programmatic buying hub when Publicis acquired Razorfish in 2009.
Fluent is led by Drew Kurth, Razorfish's CEO, Digital Marketing Platforms and Services. Kurth took some time out last week to talk about Fluent.
AdExchanger: What's the origin story for Fluent?
DK: Publicis Groupe has been looking for ways to leverage some of our experience with our clients, and existing technologies. And so, the Fluent product is based on some technology that has existed here at Razorfish and was built at Razorfish over the last six or eight years. We've built products in the past, and that's helped the agency and our clients, and then later turned out to more specifically help our corporate development.
I think also the notion that we would develop products helps to diversify our revenue stream. It's good to have other lines of business. It's certainly never going to compete with the agency services business, but it's good to explore other models. I don't think we would build software that is an order management system or something.
So the onshore team is about 10 or 12 now. It will probably go to 15. And if you add in offshore and some partners from Razorfish, people who are not dedicated to Fluent but who are working on Fluent projects, the team gets very quickly to about 60. Twenty‑something offshore, 20 full‑time Fluent employees, and about 20 or so Razorfish folks that are working on it for the foreseeable future. It's a complicated answer, right?
Slightly, but that's often the way in agency land -- nothing quite as straightforward as you would hope.
The strict answer is the Fluent payroll is 25. The offshore team is 20‑something. And then, we get some help.
You say that some of these technology products have roots going back to before Microsoft even acquired aQuantive and certainly before it sold Razorfish to Publicis. Why didn't Microsoft keep the technology bits for itself?
I'm not certain I really know what drives decision‑making in Redmond, so I'm not sure I'm the right person to answer that. But I can give you some idea of what they are and maybe that will help you understand it.
Even before Microsoft bought us, we were buying huge amounts, vast amounts of digital media for our clients. And almost part and parcel of that becomes reporting. How do we report? So way back in the day, we had some simple reporting systems that were basically hopped‑up Excel with Macros and things like that. Over time, we built these reporting platforms up into analytics platforms. But it was all around the digital media. So we've had two or three - actually maybe even four now - versions of a reporting [system] for our digital media spent for individual clients. Those pieces were in some cases co‑developed with partners, in some cases they were custom‑built for specific clients. And that's part of what we're leveraging.
Microsoft may have its own technologies they think were better than ours. Who knows why they didn't keep them? But that's what we've been talking about, is these reporting systems that have gotten more robust over time.
And then, more recently, one of the tools is something more like a year or two old, and that is focused on targeting. And so, we custom‑built an advanced targeting software application and solution for one of our clients. We are going to take that and build on it. It was custom‑built, so there's a lot of stuff that was hard‑coded specifically to that client and their needs.
But we're going to take our learnings from that and what code we can and we're going to extend that to be what we call "Fluent targeting."
So with this targeting offering, which is more recent, I can imagine some people asking or wondering, "Is this the second coming of ATOM Systems?" What would you say to them?
First of all, the technologies are super‑different. It has advanced so much in the time frame between then and now. In the marketplace, not just here at Razorfish, we've seen these big data technologies emerge. You know better than anyone how much we can do now that we couldn't do in even 2008 or 2009, and in that short period of time, Razorfish has invested in some of those big data technologies and has brought them to bear for this particular client.
And that's good. That client's happy. We're doing more processing, five terabytes of data a night for them. It's a big operation. We hope that if we can pull it out, we can add it in to the other components of the platform that being in the new platform as an ecosystem tied to the other pieces starts to make a big difference, and makes it a new value proposition to our clients.
How is Razorfish sorting out how much display media targeting and buying activity it can manage locally versus working through Vivaki Audience on Demand and the corporate hub? What's the right balance there?
For sure AOD will continue to be a partner. But what we plan to do and expect to do to abstract our technology from the demand side platforms and from the trading desks so we can integrate with any of them that our clients may have. It's great that our clients use AOD. But where they don't and they have others, we have to support them.
So what we'd like to do is pull together third‑party data and client data to advance new segmentation against them, build targeting rules against them, and then be able to publish those targeting rules not just out through the DSPs and exchanges to drive their buying, but also to CMS systems.
Our plans are focused on building those targeting rules so that we put the customer at the center across channels - so that the systems, whether they're Publicis‑owned or they're some third‑party system that your client has, we integrate with them.
Whether we're driving your site experience to change according to data and these segments of targeting rules, or display, we want to abstract that layer back to the platform so you're doing a better job of segmenting and targeting. We've done this for some clients.
Clients are all over the spectrum on how much of this they do. In some cases, our clients are already there. In others, they're just starting to go down the segmentation and targeting route. In one case of one of our clients, they were able to determine and identify 29,000 segments. Clearly, they're not going to be reacting to 29,000 segments. That doesn't make sense either. But they got it down to more like 100 or so.
And that still is a huge increase to what they had been able to do in the past, and because they're software machines driving this activity, you could actually mange experience for those 100 different segments. You sometimes don't want to pay extra money for vast amounts of creative to deal with 100 segments, and we get that. But by having them there, we found that clients are really excited about that. Because then they can start to learn more about what's working across different dimensions in ways that they weren't able to before.
Let me ask you about your title. CEO, Digital Marketing Platforms and Services. How much is platforms and how much is services?
So our vision is platforms. Services will be limited. The services will be very limited to onboarding clients on to the platform, and then support, maintenance. You've got an 800 number to call, a help desk. The field services, I think in the way that you're asking, will provided by the agencies, Digitas and Razorfish and the other agencies within the group - for example, the Razorfish consumer insights team. There's an entire discipline of 117, 118 people who are in the Razorfish consumer insights team. They are analysts and data scientists and PhDs and people who can drive into the data and build real reports and insights out of it. We're not interested in that stuff. Well, I shouldn't say that.
I'm interested. I think it's fantastic. But I'm not going to hire an army of analysts do that stuff.
Although these products have a history, you've just formally "launched" Fluent. What brought you to that point about the timing of that?
Some components are more ready than others. So if you think about a Fluent client who's come in as a "I want the whole thing," we don't have one of those. But when you think of the underlying technologies, we have somewhere between 20 and 25 clients that we're serving. So it's a pretty robust business model already. There are live clients and people are doing things. But packaging them all together, we haven't yet done. We've built this business unit to do exactly that, to bundle them together, and every part has some software development. We're wrapping them all in an interface and a tool called Marketing Central, and that has made some significant progress.
And that Marketing Central wrapper, for lack of a better word, is facing a release here in late summer, early fall. And so, we want to let people know about it.
Who do you compete with?
So, certainly we see some large consultancies getting into the space. So we know Accenture is out there in the space working. And then we know that Infosys has a platform that they've put together that's got some component or partnership with WPP. So we see those guys out there. And then, we see a vast array of one‑off‑‑or I shouldn't say one‑off‑‑targeted solutions. So targeting engines or reporting engines or analytic solutions that maybe don't cross the spectrum of digital marketing activities. And of those, you've seen the lunarscape. I've seen the LUMAscape. There's hundreds of those that have emerged over the past couple of years.
By Zach Rodgers