Taking the larger picture, how would you describe the evolution of programmatic buying globally, if you were to split it into stages?
Brian Lesser (BL): I like to think about audience buying as a broader concept rather than just programmatic. Audience buying is the notion that an advertiser can buy a specific audience regardless of the context that audience appears in. In other words, the way broadcast media is purchased and the way digital media is traditionally purchased is based on context. So, if I’m looking for sports fans then I purchase sports content and if I’m looking for the women in the audience then I purchase content that is of women’s interest. Over the last 10 years, we have developed technology that can disassociate the audience from the content.
Audience buying started 10 years ago with the concept of behavioural targeting. But in those days ad networks would define the audiences based on data. The problem was that it was an ad network’s definition of audience and not that of an advertiser. Even so, it was very beneficial and an advancement in the way digital media was bought.
The second stage of audience buying is open exchange buying or Real Time Bidding (RTB), which is the advertiser’s ability to gather data (off their site or in some cases by buying data) and define an audience and access them in a real time market place. This improves upon behavioural targeting. And you can buy the (more precisely targeted) audience on an impression-by-impression basis.
Stage three is really where the industry is starting to go in very advanced markets. Advertisers can pre-define the audiences but they buy them in private market places. As in, they have a one-to-one relationship with the publisher whereby they reserve inventory or they’ve pre-bought inventory. So they’re not buying in an open market place but reserving the inventory so that they can access those audiences in a premium environment. For our clients, that are big brand advertisers, they would prefer to buy in a private market place than in an open exchange.
In India, it’s early days and many publishers are making inventory available programmatically in the marketplace. But we still have more work to do in terms of educating publishers on how valuable programmatic can be for monetising inventory and data, and also in terms of educating advertisers about how powerful this can be to help them engage with their audiences. India is more likely going to be one of those audiences that goes from a nascent stage of programmatic to a mature market very quickly and does not necessarily have to live through the adolescence period.
What happened in early adopter markets is that publishers dove into RTB without necessarily having a proper monetisation strategy and in some cases it meant that publisher pricing declined in the short term. That phenomenon has actually scared publishers in less developed markets from wanting to move into programmatic. But India will mostly skip that stage as there are very sophisticated publishers that will move into programmatic but with a proper strategy so that they can control who buys their inventory, pricing and data.
What do you estimate is the share of programmatic buying among all digital ad buying globally? How does the number compare in (i) Apac and (ii) India? How do you see it growing?
BL: Well under 5 per cent at the moment. It would depend on whether it is digitally savvy client or not. There are some clients that we have where all of their digital is bought programmatically but such clients are in the minority. Apac is less developed from the programmatic stand point but we have global clients that have learned and developed strategies in the more advanced markets and applying those strategies in these markets.
Xaxis was launched in India in 2012. How big is Xaxis in India, and in Asia, now? In terms of (i) number of people, and (ii) volume of business?
BL: We started in India as mostly a display product using the data that was available in the market. When we started here technology wasn’t as much distributed as it could have been across our publishers. Data was difficult to extract from websites or campaigns. Coming back to today, about 40 per cent our business is now video which is a business that didn’t exist two years ago and technology is much more prevalent.
Apac is our fastest growing region by far. We manage a business that’s just under $200 million across 13 markets and have over 200 employees in the region. It is still our smallest region even when compared to Latin America, which is still emerging. However, Apac is the fastest growing as it has the benefit of knowing what the other first mover markets have gone through and correcting any of those mistakes.
There are concerns about areas like content (which are becoming extremely important) and context when it comes to automated buying. Is human intervention dispensable?
BL: To some extent human intervention has been dispensed with. It shifts what the humans are doing but it doesn’t necessarily mean that we need fewer people.
The job of a media planner or buyer will change over time and will be much less about listening to media owners pitch about audiences in a conference room. It will be more about looking at databases to understand where the audience is coming from and devising strategies for connecting with those audiences with a good creative format. The media buyer or planner of the future will be more of a statistician, analyst, engineer than they are today. That doesn’t necessarily mean there will be fewer people servicing client’s business.
Would you say the focus on ‘real time’ has enhanced the significance of programmatic buying?
BL: Definitely. We recently launched our new data management platform (DMP) Turbine in Apac and specifically in India. It has a unique ability to create audiences in real time. So as we learn more about a user we’re updating that user’s profile and re-categorising the user. What we mean by real time is when literally somebody goes to a website and puts something in their shopping cart but abandons it, we can then retarget that user on the next impression that we see. So it does remove quite a bit of latency. It also improves optimisation as the system in real time understands who is responding to what type of ad, in what time of day, in which day of the week and so on.
Michel - you commented recently that 'most clients are not capable of taking programmatic media buying in-house'? Is this is a clear ruling out of in-house programmatic for the future, at the client end? If yes, why can’t automation and systems be replicated inside the client organisation?
BL: Whenever there is a clever new medium or a buying technique, clients are interested in that and try to determine how that can affect their business. Some clients will go so far as to say that ‘I think we can bring this in-house and do it ourselves’. Having built a business over the last three years, I know how difficult it is to run a programmatic media campaign. Over time, clients will realise that they need a professional organisation to help them. Some clients will have success but a vast majority of clients will need our help.
Michel de Rijk: There will always be a handful of clients, say five or ten, who will be able to bring that in-house. You need to put your resources and funding behind it and make an interesting enough brand for people to work for. There will be a majority of advertisers who will think they can do it but not be able to do it eventually. They might not have the policy, proper funding or the right kind of talent required to do it.
You've recently mentioned in an interview that one of the company's global initiatives for 2014 is 'new products in new channels organically and via acquisition'. You've recently launched Turbine. What will these new products be like?
BL: Our product development cycle is to look at new technologies, new sources of data and inventory, and combine them into something unique in the marketplace. So a good example of a new product is Xaxis Sync which was developed first in the Netherlands and has been rolled out in four other markets. It is still in development here in India. Xaxis Sync is our ability to serve a display or video ad on your mobile that corresponds to something that is happening on your television. So you may be watching TV and a commercial comes on and you get distracted by your tablet or mobile to see a corresponding ad to what’s happening on screen.
The talent would include: audience analysts, data scientists and software engineers. What’s the split currently? Are there any specialist hubs? Is there scope for one in a specific area in India?
Atique Kazi: At this point of time, the base where we get support from is in Singapore, where a lot of technical know-how and expertise sits. Over time we look to develop these skill sets here. The employee strength for Xaxis India is nine people, and that’s more towards engaging with agencies, trying to educate them to understand the benefits of programmatic and so on.
By the end of 2014, Xaxis looks to enter 40 markets (in 34 currently). Which (8 or more) is Xaxis eyeing to move into?
BL: Middle East, where we have just gotten started. Latin America, where we have several markets live but will be looking to expand our presence further. Central and Eastern Europe, and we also recently launched in Russia. Then, we’d look at Turkey and Northern Africa.