DS: Four things: expanding the core engineering team, building the product, deploying with early customers, iterating tightly.
We’re building a very large scale data processing and predictive analytics infrastructure to capture billions of daily transactions on behalf of publishers. We process 2 billion daily events on our infrastructure. We add a terabyte daily to our 250-terabyte clearinghouse of media transaction data. We’re well on the way to aggregating a petabyte of media markets data. Thanks to the collapse in the costs for compute and storage, Metamarkets is capturing primary economic signal to publisher benefit.
What do you mean “primary economic signal to publisher benefit”?
This is data that reflects an economic transaction. It’s a principals trade – one agent in the market has something to sell, another has something they wish to buy. Each party has a set of attributes that perfectly describe the item they want to buy or sell. If we capture billions of these combinations, across geographies, buying agents, selling agents, we begin to describe an entire economic world, reflective of sentiment and activity. If we can process and store enough of this data, we can eliminate noise and locate signal. This data as recently as 9 months ago was too diffuse to capture and too expensive to store -- essentially, global media companies have been forced to behave like a national retailer who would sell billions of items, but due to an infrastructure gap, would expunge nightly all cash register and point-of-sale data, to pretty predictable effect.
Can you provide a use case?
We provide predictive analytics to a major international publisher of financial content that helps them better access real-time inventory availability across their global site network. We show them current market pricing for their inventory of selected attributes relative to comparable inventory in the market. We enable this publisher’s sales executives to understand quickly and elegantly the booking and pricing curves for different inventory tranches across multiple site zone and audience key value combinations. We graphically deliver data regarding willingness to pay among different advertisers with whom this publisher has transacted business. This is actionable data that this publisher can manually digest, or may route automatically to allocate inventory to capture margin upside. We are working to help this publisher to dynamically price inventory to cross with otherwise-unmet demand. We store historical transaction and order data at scale, enabling this publisher group to access almost 10 years’ worth of their own monetization history in order to understand quickly and seamlessly inventory category and advertiser category trends. For participating publishers on our platform, we provide proprietary market data products such that publishers find opportunity in market trends for the first time.
What problem is Metamarkets solving?
We help global media companies deal with the data fog and confusion engendered by high velocity, shifting markets. Publishers can’t respond to what they can’t see – there is real power in capturing data at scale and providing low-latency, highly-actionable business intelligence. This is a entirely new generation of business analytics for media companies, powered by bleeding-edge infrastructure and an industry-leading clearinghouse of transaction and order data. Metamarkets helps global media companies derive true value from their data exhaust, which is nothing other than the equivalent of other industries’ cash register, POS, and shipment data. This is real data, reflecting real dollars, and it fuels our analytics and market data products. On top of our data capture infrastructure, we perform stream processing, post-processing and predictive analytics. We build dynamic predictive models, we syndicate market benchmarks and indexes. My co-founder Mike Driscoll has led the development of a stunning real-time graphical interface to viscerally display trend and uncover opportunity. Transaction data is captured, staged and analyzed in a secure and redundant hosted environment; our analytics and visualization are delivered via the browser, for the benefit of global media companies. Publishers need build, deploy or maintain nothing. (They do however need to remember their platform log-in credentials, though we do reset passwords when folks ask very nicely.)
As with any sophisticated platform, we have multiple infrastructure projects underway at once -- and things are only getting better for our customers and partners. Mike and I knew we were setting out first and foremost to build a technology company to fill the crucial gaps in publisher monetization infrastructure. Our critical differentiators are the team and technology we’ve been assembling. We’ve recruited from Google, Harvard, Stanford, LinkedIn. We co-host events at TechCrunch and are sponsors of the esteemed Data Drinking Group (DDG) here in San Francisco. It’s a fantastic team, part of the larger data hacker community springing out of the startups here around South Park. We’re having fun and building a killer product, though we do find ourselves eating quite a few late night Thai and Pakistani meals at the office.
Any change in the problem which Metamarkets said it was solving in April?
No, only increased focus as a result of validation of our alpha product. Our goal remains what it’s been since we sketched our initial prototypes last year: to enable global media companies to effectively analyze, price and predict the value of their inventory. We’ve seen fantastic uptake and great results with large media suppliers, regardless of inventory class. In the case of one large multi-domain publishing group, we’ve seen aggregate pricing improve well over 25% since deploying our solution across multiple classes of inventory.
We suspected that the media markets were rapidly collapsing upon each other, to the point where distinctions between ‘spot’ vs. ‘upfront’ vs. ‘futures’ were being rendered less meaningful. What we’ve been surprised at is the volatility and dynamism that exists in the electronic media markets. We are heading for a single, electronic, real-time media market. On the publisher side, we’re seeing collapsing premiums for reserved inventory tranches. In certain geographic and contextual super-categories, contangoed futures spreads are tightening or in some cases going into backwardation. Traditional pricing and allocation models are being upended. To quote Dr. Peter Venkman, “Dogs and cats living together! Mass hysteria!” The point is, publishers have until recently not had the data or analytics partners to begin to grasp these market dynamics, let alone exploit them to revenue and margin advantage. We hope to help.
Where do you see the Metamarkets toolset fitting into the ecosystem? Is it a yield optimization tool, data management platform.. an acronym of some sort?
I will defer answering for fear of suddenly showing up on the Kawaja bubble chart. It’s a point of honor that we’ve avoided showing up on that chart. We haven’t been slotted anywhere. Maybe it reflects our orthogonal approach, or the fact that we’re outside the echo chamber here in San Francisco? Regardless, we’re free to build world-class infrastructure to address publisher pain. Acronyms just aren’t helpful anymore. They’ve lost their meaning, become the punchlines to jokes. When there is a real revenue crisis underway among the world’s most esteemed media brands, I’m not sure an acronym is going to do much. I’ve been in this industry long enough to know that there is more than a fair amount of cynical marketing behind much of what goes on.
Metamarkets is a data-heavy publisher-side technology company. We are 90% engineers. Given the enormity of the technical challenge and market opportunity, that feels about right. If there is critical headcount being deployed toward marketing or webinars, that suggests to me a great deal of effort is going into the manufacturing of a problem on the part of the marketer, in the hopes of catching some percentage of the incipient demand the marketing has created. From our perspective, the problem is glaring: data overwhelms media suppliers, over-investment in demand-side technologies has only accelerated the deepening of pricing craters, and publishers are getting creamed. I’m not sure we need a webinar for that. Metamarkets is tackling what I call ‘first order’ problems: price discovery, inferential and predictive analytics, dynamic allocation, market indexing, risk management. If we need to be slotted somewhere, put us in the ‘For Publishers Only’ category. I’ll leave the pigeon-holing to others.
Are you in-market yet? And, what is the target market for Metamarkets?
Yes, we are in market with six alpha customers within our target market of large, integrated global media companies. The requirement for working with us is that we share a vision of how the electronic media markets are evolving: toward a single, dynamically allocated sales channel. For a pub to meet the future, there is some pretty heavy duty infrastructure required, backed up by immense data sets and analytical horsepower.
To this end we are fully integrated with the major ad serving technologies from OpenX, ADTECH and DoubleClick, in addition to custom ad servers that large publishers have developed in-house.
Where does your solution fit within the real-time ad world such as real-time bidding?
The consistent narrative will continue to feature ‘publisher as price-taker rather than price-maker.’ It’s all well and good for a publisher to make inventory real-time biddable, but how in the world does a publisher set about determining fair value for tranches of inventory across multiple, rapidly moving marketplaces? Without the ability to stream process and analyze these events at massive scale, a publisher is bringing a knife to a gunfight. There is no doubt the market is moving toward a real-time protocol. But from the publisher’s perspective, this should not be about simply putting inventory up for real-time bidding. That’s the scenario where the publisher is price-taker. In the version of the story where the publisher is price-maker, data and analytics enable the publisher to make inventory real-time sellable. Perhaps that’s the acronym right there: RTS. Metamarkets makes publisher inventory available for RTS, real-time selling.
Is there an opportunity for the demand-side with Metamarkets?
No. We are focused on the publisher side only. Publishers have been taking it on the chin for years.
A year from now, what milestones would you like the company to have achieved?
In surfer/NorCal/start-up parlance, we will continue to ‘kill it’ on behalf of global media companies. Metamarkets visualizations, predictive analytics and market data will be running on global media ad ops, monetization analytics, sales and CXO screens all over North America, Asia and Europe. We know it’s a big plan, but if the problem and concomitant opportunity are as immense as everyone knows them to be, half-measures simply won’t do, will they?