On Marketing Technology

MTS: Tell us a little bit about your role at Metamarkets and how you got here?
The origins of Metamarkets stem from my early days as a data scientist, entrepreneur, and academic. I began my career as a software developer working on the Human Genome Project, later founded an early custom apparel company called CustomInk.com, and left to pursue a Ph.D. in computer science and biology. In those roles and others, I spent many nights crunching through big sets of unstructured data, trying to discover an insight. It would take hours to generate a trend line visualizing a few million data points, and I felt I was spending the majority my time preparing and structuring data, versus analyzing it.

When I came to Silicon Valley, my dream was to build a product that took the drudgery out of doing data science at scale, and made data accessible, interactive, and visual for regular people — not just developers. I felt that too many existing dashboard applications were not flexible enough to support the kind of drill-down that enables ad hoc discovery. I founded Metamarkets to fill that void.

We quickly discovered that the digital marketing world was a great fit with this tool. Programmatic marketing is arguably the world’s most data-intensive vertical, with daily transaction volumes — across channels that include Google, Facebook, and thousands of publishers — that are 100X greater than the NY Stock Exchange. Marketers were hungry for tools slice and dice through audience segments and campaign metrics across channels, without submitting report requests to a data science team. Over the last six years, Metamarkets has grown to become an industry standard in interactive analytics for programmatic marketing.

MTS: What are the technology trends, challenges, and developments that are going to impact the programmatic marketing sector?
The biggest development that’s going to impact the programmatic sector is the convergence of martech and adtech. We’re starting to see marketers getting more hands on and involved with programmatic technologies. As that happens, the language of marketing will start to permeate and eventually dominate the language of ad tech. So, for example, ad tech teams talk about impressions, auctions, bid requests and bid responses – but marketers talk about people, campaigns and actions. The martech language is likely to be what ad tech adopts as part of this convergence.

The biggest challenge is that of the tension between cooperation and defection in the industry. Marketers want to see cooperation between the major ad tech channels, in terms of standardizing metrics and sharing data about audiences and performance. But unfortunately, the biggest players, namely the duopoly of Facebook and Google, have a plenty of incentives to defect and operate by their own rules. That’s a significant challenge for the long-term success of programmatic, but marketers will play a key role in demanding cooperation, as they are the ones writing the checks that these channels need to survive.

One of the biggest trends this year is the push marketers are making for more transparency into the media supply chain this year. There’s a critical need to unify and consolidate data from disparate channels into a single source of truth. I believe that ultimately, transformative transparency will require transformative technology. If marketers really want to own that transparency and get event-level data about their campaigns, they’ll have to make significant investments in infrastructure because of the unique scale and complexity of programmatic data. Marketers are going to have to make a number of important choices on what they build or buy to create the next generation marketing stack for their organization.

MTS: Could you explain the link between chaos and innovation?
I believe chaos is one of several key ingredients to foster innovation among engineers – it is a key pillar of my engineering management strategy. Too much structure can be stifling to a merry band of rebels conspiring on the next new thing. Whether through allowing engineers 20 percent of their time to work on unorthodox projects, hosting week-long office hackathons or giving a helpful nudge to an internal skunkworks initiative, innovation breeds best in a bit of chaos.

MTS: How does Metamarkets look to leverage its technology for its clients?
The value that our technology provides for our clients is all-around transparency. When data gets big, it often gets difficult to make sense of – that’s true in every industry but especially in advertising and marketing. Metamarkets has spent the last several years developing a full-stack solution that allows the operators of programmatic marketplaces and buyers of programmatic inventory on those markets to get radical transparency into the trillions of impressions and billions of audiences that exist throughout the programmatic ecosystem.

Why does that matter? We have discovered that programmatic marketplaces that are the most transparent are also the most successful. Market transparency truly drives market activity in this industry; when internal users understand what’s happening, they are able to optimize their marketplace strategy. And when buyers of programmatic inventory have a clearer view into what they are buying, they buy more effectively and they buy more. That trend extends upward through the supply chain – once marketers have the transparency they seek, it will make them much more effective at their jobs.

MTS: Could you give a brief overview of Metamarkets’ full-stack, SaaS solution?
Marketers need access to the freshest data possible – it’s really hard for business users to make timely decisions if they can’t get answers to their own questions. And if data scientists have to be called in to answer questions, it slows down the process and increases the barriers to getting insights.

Metamarkets solve that problem by providing an intuitive, point-and-click interface for interactive analytics that turns each user into his or her own data scientist, with the power to explore all corners of the data. Programmatic buyers and sellers use the platform for price and inventory discovery, reach and frequency planning, and many other insights into the media marketplaces where they operate. As the only analytics platform purpose built for the unique complexities of the programmatic space, Metamarkets plays a critical role in helping the world’s top programmatic marketplaces and buyers turn mountains of data into revenue-driving insights.

MTS: What startups are you watching/keen on right now?
I’m a big fan of the growing API-driven economy and the companies that are part of that, both in marketing and other industries. We know there’s a great future in artificial intelligence and the optimizations that are possible with greater algorithms, but before we can get there we need to actually build the data foundation for that intelligence.

The prelude to intelligence is automation, and the primary driver of software automation and more frictionless data flows are APIs. So companies like mParticle and Segment.io in marketing, Twilio for telcos, Braintree and Stripe in payments – these companies are quietly and solidly building the software plumbing that all of the intelligence built in the future will stand on.

MTS: What tools do you believe make up the ideal marketing stack in 2017?
The way I think about the marketing stack is as layers of tools for data acquisition, data storage and data visualization. A user interface that marketers have as part of their daily toolset is only as powerful as the database and connectors underneath it.

On the data acquisition front, many marketers are coming to realize the challenges of connecting to the data within all of their channels. So marketers must first find tools that help them effectively and cleanly acquire data from their channels. Often times this is something that marketers need to build themselves with a team of data engineers.

The next layer up the stack must help find a place to store all that data that you’ve acquired. In this case, there’s a new generation of databases that are up to the task in terms of managing the unique scale of digital media and advertising data. Large scale database tools exist in the cloud like Amazon’s Redshift and Google’s BigQuery, and there are other proprietary tools sold separately like Snowflake. At Metamarkets we’re proud to have developed Druid, a large-scale open source database tool that has been widely adopted in several industries.

The most important element of this database tier for marketers is that it is fast, because that directly powers the third layer of the marketing stack: a visualization interface layer for business users. Once an analyst has acquired and stored their data, the quality they need most is interactivity to ask questions and get answers at the click of a button.

I believe the future of marketing technology is real-time; not every channel is ready to go real-time today, but eventually in the next 3-5 years, all marketing channels will provide real-time data.  For those marketers looking to future-proof their stack in 2017, they ought to choose tools at each layer that are able to work with real-time data.

MTS: What is the essence of a standout digital campaign?
The essence of a standout digital campaign is audience – you want to be able to hit the audience you want to hit. Ultimately, the success of audience targeting all comes down to the quality of data on two sides. First, quality of 1st party audience target data you bring to the table, as a marketers, and the ability to wrangle and manager your own CRM data effectively. And second, is being able to work with channels that can reliably and deterministically map your target audiences to their audiences.

MTS: How do you prepare for an AI-centric world as a business leader?
Unfortunately too many business leaders have been caught up in the buzz of AI, and in their desire to embrace AI they’ve skipped an important first step: unifying and harmonizing their data assets. They need to focus on data first, intelligence second.

Simply put, you cannot build intelligence on an unstructured data lake: it will sink.  The most important way to prepare for an AI-centric world is to focus on developing a very clean high-integrity data foundation across departments. Few organizations do this well, and marketing companies are no exception. The companies who are most successful in AI so far are companies like Google and Facebook who are by all accounts a decade ahead of the field in building incredibly powerful data foundations.

This Is How I Work

MTS: One word that best describes how you work.
Trust.

MTS: What apps/software/tools can’t you live without?
Evernote – it’s my amnesia anecdote machine.

MTS: What’s your smartest work related shortcut or productivity hack?
Cancelling meetings. Every Sunday night, review your calendar and cancel any non-important meetings for the week, but don’t replace them.  I get my best thinking during those freed-up hours.

MTS: (What do you read, and how do you consume information?)
I’m currently reading Nathaniel Popper’s Digital Gold, a history of Bitcoin, because I think that cryptocurrencies have implications that go far beyond just currency.

I read voraciously on my Kindle and iPhone 7S, mostly non-fiction, and try to stay off social media.

MTS:What’s the best advice you’ve ever received?
Marry your best friend.

MTS:Something you do better than others – the secret of your success?

Rational exuberance and optimism. The ability to always look forward and see the positive side of any turn of events, because ultimately the future belongs to the optimists.

MTS:Tag the one person whose answers to these questions you would love to read:
Shashi Upadhyay, CEO/Founder, Lattice Engines.

MTS:MTS: Thank you Michael! That was fun and hope to see you back on MarTech Series soon.