Proximic Bringing Contextual Search Technology To Display Says CEO Pieper

By
  • Facebook
  • Google Plus
  • Twitter
  • LinkedIn

ProximicPhilipp Pieper is CEO of Proximic, a contextual technology company, which announced today that it was integrating its technology with real-time ad platform, AppNexus. Read the release.

AdExchanger.com: How has Proximic evolved over the past few years?

PP: Proximic is built on a unique search technology foundation, designed for high performance and precision contextual matching. Soon after developing the initial platform, we quickly identified the strong need for this advanced technology within the online advertising space. In contrast to other technologies, our methodology provides a higher level of flexibility and performance that allowed us to offer more precise, flexible, scalable, and effective solutions for our customers.

Looking back to 2009 when the U.S. advertising market was going sideways, the language independence of our platform enabled us to enter the Chinese market. Focusing on solutions for display and search, we successfully navigated the tremendous growth of the Chinese online market dominated by highly dynamic content, a challenging IP infrastructure, and low valued online ads.

When we entered the U.S. display ad market earlier this year, we already had cut our teeth within the huge data environments abroad, while competitors were still making their platforms RTB-ready back here in the States. We built our data deliveries based on experience we gained working with latency-sensitive large players in real-time.

Why position your company's product as "rich data for digital media trading"?

Monetization of inventory through exchanges and in RTB is a market reality and we have no doubt, that it dominates standardized/non-direct ad buys. Reliable and impactful data is essential in this context. It helps buyers make sense of the impression before placing their bids.

Proximic provides media traders with rich and precise data through our integrated technology platform that combines Contextual Data (to identify website content), Brand Protection Data (to add confidence in ad placement) and Audience Interest Data (to identify specific audience interests). This combination allows for better targeted ads and filtering of content across highly granular topical categories. As a result, our data empowers media traders to make more informed targeting decisions, allowing campaigns to run more effective by identifying premium ad placements, both with confidence and at scale.

What is Proximic's competitive set today? How do you differentiate?

Across the display ad ecosystem there are two groups, those performing media transactions and vendors supporting the media buying. Proximic resides in the support group competing with everyone else who is selling value-added services, tools, and data improving the process of media buying. As we serve different functions around page contextualization, brand protection and audience interest, there’s no specific vendor in the market we directly compete with.

We are focused on applying advanced contextual matching technologies to build unique data and analytics solutions. Unlike other approaches that are based on static set of rules or fixed taxonomies, Proximic works dynamically matching against a self-maintaining and continually changing set of hundreds of thousands of categories. We overlay that granularity level with the data our customers want to target against – whether against categories or keywords – making it directly actionable for our partners.

With all the noise in the vendor space, partners are becoming increasingly focused on 3rd party relationships that give them increased leverage. Our value proposition is compelling as we give partners a fresher and more transparent real-time inventory assessment, better customization around targeting data, and the ability to execute their businesses globally.

Given the current regulatory discussions under way both in the U.S. and Europe in regards to online behavioral advertising, a do-not-track list or any limit on cookie data is good for Proximic, correct?

This is a hot topic right now. There is no easy answer to the trade-off between privacy concerns and the value which behavioral data brings to online ads. End consumers certainly don’t understand the subtle difference between right and wrong business practices (even within our industry there are a lot opinions on the shades of gray). With the discussion around do-not-track becoming more visible by the day, we have to anticipate the possibility of imposed data limitations. Many players in the ecosystem will consequently have to take a serious look at how they run their business practices.

I’ve read many statements lately, saying that limiting tracking will “break the online model” and sites will have less advertising revenues. Actually, those who consider the absence of behavioral targeting as less accurate are misinformed. There is a lot that can be done with analyzing content. Advertisers will continue to reach their audience, only the methods will change. Therefore, we can foresee already that regulation will certainly strengthen Proximic’s Contextual Data business.

But even beyond that I believe the same will apply to our Audience Data business. Let me give some details on how Proximic handles data. First of all, we generate audience interest profiles based-on data originating from our partners. Secondly, we don’t use any personally identifiable data. Instead, we look for contextual signals on the web pages a user visits. Using data that cannot be linked to explicit PII behavioral targeting becomes privacy non-intrusive, which works equally as well as other methods. That’s why we believe “do-not-track” by default is not necessarily the right answer.

Can audience interest and intent data be extracted from contextual/semantic data? if so, how?

Absolutely. At Proximic we analyze contextual signals from the web pages a user visits within our partner’s network. Within that stream of data we have the ability to identify clusters based on signal density. We relate these clusters with topical segments, for which the user has been showing interest. Observing segment occurrence over time, allows us to identify profiles of change in interest. These rich profiles present a powerful basis for optimization and can result in the similar impact as intent data. That said, there are a couple of underlying preconditions that need to be in place to do this effectively. First, it’s essential to have enough signal granularity, so that profiles sufficiently differentiate. Secondly, there needs to be a high degree of processing efficiency and scale.

I would also like to add that intent data gathered from data sellers can be great, but can also be spread too thinly across a client’s inventory, which may lead to a below average effect. By doing a better job at refining data broadly, which a partner already has access to, we can substantially reduce data costs by taking the data seller out of the equation. Plus by “right-sizing” the cost for data it helps to avoid artificial and unhealthy inflation of media cost.

What is your target market from a buyer, seller and vendor perspective?

We’ve developed our solutions to be relevant across the ecosystem. We provide our services to the entire spectrum: Agency Trading Desks, DSPs, Exchanges, SSPs, Publishers and partially to other vendors. Our customers use our data with very different purposes:

With demand partners we tie into their real-time bidding processes by determining the value and safety of a page and making users targetable in a privacy safe way.. This makes inventory more transparent and gives buyers greater control and effectiveness.

For sell–side partners, our data helps increase the inventory’s value across different sales channels. Firstly, direct sales can be supported with better targeting and advertisers can be more confident of their placements. Secondly, with our data, unsold, but clean inventory can trustfully be passed into the RTB channel. Lastly, we help dynamically monitor inventory subsets, that based on their quality, should be routed to other advertising channels.

Do you think large publishers effectively leverage data today? What can they be doing better?

Not entirely. Large publishers have a lot of data, but so far, have not reached the sophisticated levels of data processing that ad networks are currently employing. The easy answer has been to either leave the monetization of remnant inventory to networks, which led to discussions around data usage, or shut down access to remnant inventory entirely and leave money on the table. Both not entirely satisfying solutions.

We can expect this to change on two fronts. First, the tools for analysis of “owned and operated data” for direct sales will improve with solutions such as those Proximic can offer. Secondly, solutions that improve the exchange and control of publisher data for the monetization of remnant inventory will come to market (DMPs are good examples). I am confident next year we’re going to see very interesting developments in both aspects.

What happened with your deals in 2008 to syndicate product listings from eBay, Shopping.com and Yahoo!’s Shopping Network and creating a contextual ad offering? Any key learnings you can share?

In getting to where we are today, we engaged in various use cases that uniquely demonstrated the abilities of our technology. In working with affiliate listings we were one of the first companies capable of taking tens of millions of shopping items and purely based on matching the associated text delivered significantly more relevant matches to a publisher page. However, along the way we found our solutions for the display market to be a lot more impactful and valuable.

These deals allowed us to gain an immense amount experience around keywords and mid/long-tail sites that are fundamental to our current platform. I am certain we will continue to benefit from these experiences as we engage with dynamic creative and optimization technologies serving the need for more performance in display.

How do exchanges and real-time bidding impact your business?

It’s a driver for our business. Real-time bidding is certainly emerging as one of the more impactful ways to monetize impressions for standardized buys. Projections on the growth on this channel vary; however, it is clear that the current volumes of impressions are creating technology challenges for media traders. As an example, the amount of inventory that is theoretically accessible to the demand side is breathtaking. In making sense of that inventory, the processes of data analytics are far too costly for individual buyers and repetitive across all participants. The prohibitive costs and inefficiencies around this type of data call for solutions from vendors like Proximic.

Will you need new funding?

We are currently not raising funds. At this point we’re focused on building-out our market share and ramping-up our sales, marketing, and development efforts and with all the market dynamics in play, we continue to keep an eye out for opportunities. We therefore continually talk with investors to see whether more capital can help us grow faster.

How many employees in the company today? A year from now, what milestones would you like the company to have accomplished?

The Proximic team today consists of 17 very talented individuals.

With Proximic, we have the potential to play a crucial role in improving the effectiveness, confidence, and trust in display ad placement. A year from now, we see ourselves as the leading, trusted reference source for value-impacting data improving the transactional decision making for media traders by facilitating data between publishers and advertisers. In a way, becoming the “The Bloomberg of the Online Advertising Ecosystem".

Follow Proximic (@Proximic) and AdExchanger.com (@adxchanger) on Twitter.

  • Facebook
  • Google Plus
  • Twitter
  • LinkedIn

Email This Post Email This Post

By on at

Leave a Reply