Having come a long way from its mobile video days, Waltham, Massachussetts-based Nexage appears ready for the next stage of its real-time bidding (RTB), mobile ad exchange strategy given the recent round of investment led by telecom SingTel’s venture group, SingTel Innov8.
Though Nexage CEO Ernie Cormier cautions the investment is financial versus strategic and SingTel has said there are “no strings attached,” the Singapore-based telecom is clearly getting into the mobile ad tech game – also indicated by its March acquisition of Amobee. In fact, Nexage has agreed to expand into Singapore and Asia as part of the deal.
Nexage’s Cormier discussed the details, industry trends, and the company’s roadmap ahead, with AdExchanger…
ERNIE CORMIER: We’ve always looked at Asia and, frankly, Europe as the next two candidates for expansion and you want to do it with partners on the ground. SingTel has made it clear that they will work with us and provide that partnership on the ground. Above and beyond that, SingTel has a great ambition to move beyond the core telecom space. Today, they’re doing work in addressable TV and mobile advertising – which is one of the areas that they identified as a core value-add service opportunity.
The investment into us, I’m sure, was part of that ambition, but they recognized that we as an ad exchange could be the foundation of a business moving forward. Nexage sits in the middle of the transaction. Singtel came to the same conclusion – that an ad exchange really is the center of the hub and the place to start it you want to build something with a lot of value that could touch basically everything in the business as opposed to more peripheral business opportunities.
That said, we see the ecosystem evolving. We see a place for the carriers and telecoms, yet they’ve lost the ability to set the pace – and set the rules. The business has evolved well beyond the carriers, so we’ll work with carriers on an opportunistic basis, we’ll work with SingTel, but, ultimately, it’s not a key part of what we’re doing here.
It depends on which sector you mean. We work across large branded publishers as well as very large app developers and we also have smaller, more traditional game publishers and developers.
On the large, branded publisher side, we’ll compete with companies like AdMarvel, Mocean, but then again we also partner with those companies very extensively. On the pure gaming side, our competition is around smaller companies coming out of the ad serving space, some have some exchange capabilities, companies like [Velti’s] Mobclix or Burstly.
The way your exchange technology works, it sounds like it’s not one product fits all. You have to create several. Can you expand on that a little bit?
If we go at it from the technology or product perspective, we view our core asset as a platform upon which we build other products, and you can even call them services if you want, but they are not services in the traditional sense.
We have a core platform upon which we have built mediation capabilities, or so called yield optimization. On that same core platform we’ve built a true real-time bidding exchange. We think we’ll be the future of a large part of mobile advertising and become that way for display.
Looking at the difference in RTB implementations for PC-based display and mobile, how has that gap closed or evolved over the past year or two?
The answers play off the complexity of mobile and the differences between mobile and display – and it has far reaching impact.
In [PC-based] display, real-time bidded exchanges have become to some publishers, anyway, the land of random inventory where impressions go to die. There is plenty of data available in display on the buy side, but publishers have lost control of their user data and did several years ago because of the use of third party cookies and data exchanges came along.
Because of the limitations with mobile, where you do not have the ability to use third-party cookies or anything even close to the functionality of online, data for targeting and mobile really has to come from the supply side. You can link third-party data in, but that also has to be done on the supply side. What that calls out is a need in the marketplace for a place where buyers perform discovery across a large swath of impressions, and do targeting against those impressions. But it all has to be done at the point of purchase. It can’t be done basically after a purchase by going to an ad exchange and linking up data to impressions that you’ve bought in bulk at very low prices. This calls out for a real-time bidding exchange, which is what we’ve built.
We designed the data structures around the bid requests, which ended up becoming an open RTB mobile standard. We designed it knowing the role of data in mobile and how it comes from the supply side, and that way when an impression gets presented on that exchange. The demand-side platform or buyer can scan across thousands of impressions simultaneously to pick out the impressions that they want, get the data they need, to first of all determine that they want it, and secondly price it and get those impressions.
What are the challenges that you’re seeing with mobile DSPs today? For example, it would seem to me that the PC is more real-time than mobile given the connection that is necessary to a mobile device.
I’m not sure I would see the distinction there. Remember, the only place that really becomes relevant, as I understand the question, is on the recent question of location. Some buyers want location that is literally just minutes old and depending upon the architecture of how you’re delivering the advertising, because of caching, you have to learn to work with that or work around it; it’s just one of the learnings that we’ve applied over the past six to nine months.
What about the latency issue? Is it possible to target mobile cookies with little latency?
Latency is latency and you build for it, monitor and test it. If you’ve got a latency issue with an ad network over time, at least in terms of our platform, we stop pinging that ad network. But in terms of any given ad impression, if you pop open your phone and go to an app or mobile website and an ad is requested for it, we’ve already engineered the latency issues into a manageable degree so that we can handle it. There is no issue there and it doesn’t relate at all to the question of cookies or not.
Our auction is real-time, there is no difference between what our auction is doing and the display auction in terms of whether it is real-time or not. You are being presented with an opportunity to buy an impression and when you win that impression, that ad is getting shown. We do use some caching and buffering to handle out-of-norm latencies and things like that, but that’s more of an edge case than a common use case.
One thing we have noticed with the mobile ‘native’ DSP’s, and for that matter the online DSPs, is that they are strengthening their algorithms at a fairly constant basis. We’re seeing that in their ability to target on context and using first-party data, and we see that in their bidding behavior. One thing that leaves us very bullish on RTB is how well they move their algorithms forward; whether they are going after the performance spend or after the brand spend and whether they come in from mobile or online – that is something that bodes well for the industry and for RTB as a key technology.
What’s the most popular way right now that mobile DSP’s are targeting users?
It’s an interesting question because we often talk about how we can look at an impression on an RTB exchange and there will be multiple bidders going after it. We’re seeing some of the impressions getting a bid density of well over a couple dozen bidders and growing. One set of bidders will be going after the impression because it has an ID and they want to do CPI (Cost Per Install) with it, and it’s got also the platform that they are targeting, let’s say it’s iOS North America. Another set of bidders is going after it because it’s got location data they want, so they’re bidding on it for location. And then another set of bidders can be going after that impression because it has some demographics associated with it.
Thinking about the mobile auction environment, how critical is device ID in all this? I’m referencing, in part, Apple’s deprecation of UDID.
The device ID is important not so much for the exchange per se, but to a certain type of business within a mobile advertising, and that tends to be certainly anyone doing a CPI or a CPA (Cost Per Acquisition) campaign. In that case the ID is used for conversion tracking. What that tells you is that many of the substitute technologies for IDs that are out there – and we’re supporting major ones – can be used for conversion tracking.
So deprecation, while it’s a disruption and it causes an uncertainty, it’s probably not going to have any long-term damage to the business. UDID was never supposed to be used as a mobile cookie, but that doesn’t mean that it wasn’t being used. Deprecation of UDID will obviously cause a problem for buyers who wanted to build profiles, which arguably was not really the intent for the UDID to begin with.
What’s going to be the impact of the cross-device, cross-platform login? Is this a replacement for the device ID?
Even before Apple deprecated the UDID, and to your point of publishers working across both display, online and mobile, if publishers are in a world where the very use of their app or website encourages, if not requires, login, then they are now in a world where they can replicate the functionality of the third party cookie, but within their universe they can still control that data.
To a large extent that is the Holy Grail for a publisher – to be cross-platform and have a functionality that encourages, but doesn’t require a log in. You now have the ability to control and gather your user data, and provide a single point of opt-in, or at the very least opt-out for your customers. You can see that there are publishers out there that have a very advantageous setup where they’re linked across display and mobile and they’re ability to target in a way that still lets them control the data and not lose control the way it happens in online display.
What’s happening with mobile creative through your technology platform today and where do you see that going?
Creative is one of the areas that we think is ripe for improvement. There are a few companies that are doing some work in that space, but the industry needs to advance the creative. However, the state of mobile creative shouldn’t be all that surprising. If we place mobile advertising in the historical context of where display was in ’97, display ads were a little embarrassing if you looked at them today.
People have to remember that’s fifteen years of evolution. Mobile is not going to take fifteen years. It will be much faster, but we still have a lot of work to do to improve the creative. What we’re seeing today is interesting use of interstitials with video or rich media, and then within an app you’re seeing more of what could be called product placement – in-app advertising or integrated app advertising, which is still in its infancy and requires way too much production and code involvement for it to be scaled right now – it’s ‘early days.’
Partners like Celtra are leading the charge with rich media. Nexage continues to expand its capabilities on the exchange and support the new ad units. It’s not difficult for us to add them to our platform.
Can you talk about run rate through your technology infrastructure and shed some light on profitability?
Unfortunately, as I’m sure you could have anticipated, our investors really don’t let us talk about financial metrics. In terms of volume metrics we’re running right now at about twelve to thirteen billion impressions a month and that’s growing. The volume through the exchange is growing. Mediation is growing as well. Our revenue is increasing month over month. We’re seeing great growth.
Can you explain about what you mean when you say mediation is growing?
So, you have real-time bidding, which is where the true auction is taking place. For the most part, those are DSPs taking part in the auction and numbering around fifty today and growing rapidly. Those are both mobile and online DSPs. And, there are a couple of online ad networks that have real-time bidding platforms as well.
Then you’ve got the more traditional mediation of yield optimization platforms or products – however you want to refer to it. This is the older platform we’ve had for over a couple of years now and comes from the world of SSPs. All the major ad networks, certainly all of the ones in North America, Western Europe and ad networks from around the world, the JumpTaps, Millennials, etc. – these guys are all plugged into our mediation platform. They’re not bidding on inventory per se. What’s happening is that our algorithms are keeping track of who’s buying right now and who’s paying what, and we optimize whom we offer an impression based upon those two variables. We also do country-by-country optimizations as well. It’s not that the ad networks are bidding, it’s that we’re keeping track of what they’re buying and what they’re paying and we’re offering them impressions optimized for the revenue.
How many people is Nexage today? Expansion plans?
We’re at about 42 people and based in Waltham, Mass, but at the end of the year we are moving into Boston. We’ve got a sales office in San Francisco and in New York City; and we’ll be expanding both of those with more sales people.
Finally, in the next 18 to 24 months what do you think some of the key milestones that you want Nexage to accomplish?
We have metrics internally that we need to hit and will. We also see real-time bidding (RTB) continuing to become a more and more relevant part of the mix of programmatic buying and mobile.
One of my personal goals is to accelerate the growth of RTB within mobile such that we hit that 40% mark way before [PC-based] display did in terms of the total length of time.
You’re going to get to a point in time where it’s probably true that rich media constitutes somewhere between 15-20% of the total ad units, and when it climbs over 50%, you will see the value of the ad go way up relative to both the consumer and the advertiser. That’s when brand spend will have really arrived.
By John Ebbert