Platform vs. Product: Why Data and Inventory-Agnostic Design Limits Programmatic


‘Programmatic’ thrives on the notion that the historically inefficient advertising industry is democratizing through the real-time exchange of ad budget for the instantaneous intersection of inventory and data.

As with any efficient market, pricing in programmatic would thrive on perfect information. However, with unreliable data and publishers of all measures of quality, the market for real-time advertising is far from efficient.

Even as the industry matures, programmatic remains less automated than many suppose – many transactions executed programmatically are still preceded by a very human negotiation.

One thing is for sure: ‘programmatic’ is too promising to languish at a standstill. But is the widely popular ‘platform’ approach to progress, wherein largely generic DSP execute programmatic exchange differentiating only in vague terms, the right one? One could argue that indeed gravitation to commoditization is truly the only way to achieve an efficient programmatic market. But in an industry yearning for profits, and slowing progress towards programmatic perfection despite the platform approach, one must consider the direct opposite as a viable strategy for industry players at large – a move towards the productization of programmatic, trading openness for closed and generic for proprietary.

If the attributes that define the value of programmatic inventory plus data were more clearly valued, we could enjoy an efficient commodity market. In that case, we would arrive at a few simple exchanges, much like NASDAQ and NYSE dominate equity trading. But with disparate data and attributes which hold different value to different buyers, such a future is unlikely to be delivered profitably. So why is our industry trying to achieve it?

Perhaps it’s a vision for a winner-takes-all outcome, where the leaders in neutrality and transparency gobble up the industry. But even in this scenario, profitability and long-term sustainability would be dubious for any commercial provider.

Product Creates Profit

The opposite of the status quo via ‘platform’ is ‘product’ and it is the only route towards progress and profit. With a product, providers shun pure openness and compatibility for uniqueness and proprietary. With a product-driven approach, differentiation takes priority over commoditization and unique proprietary design fuels unique deliverables which advertisers will be willing to pay a premium for. In an industry that is constantly benchmarking, it is possible that an alternative approach – pure differentiation could be the way to break away from the pack and claim market share.

What Defines ‘Product’?

Many of the platform concepts alluded to in this post have product-like features, such as unique algorithms, user-interfaces and more, but they can’t get past the openness and interoperability that could lead to their demise. They seek to remain transparent and agnostic, yet simultaneously claim uniqueness. This is the programmatic paradox.

True product, and true uniqueness is just that – it’s truly different. As Peter Thiel said in his best selling book ‘Zero to One’, competition is not always good- that a business should be so uniquely good at what it does that there is no substitute. This type of approach is a major contrast to the look-alike approach in the programmatic business, and we’re not talking about audience modeling.

How Can Programmatic Productize?

Programmatic can only really differentiate on a limited number of fronts, namely ‘process’, data and inventory. And whereas process can always be improved, and inventory will often go to the highest bidder, better data is the king of all deliverables – the most exclusive of any proprietary offering and the answer for platforms making the shift to product. Data is finite, data holds intrinsic value and without data, programmatic lacks the third dimension, transacting only on the availability of inventory, without knowledge of its audience. And yet too many programmatic providers pseudo-differentiate towards better data with attempts at a better process – more efficient, more highly targeted or both.

Meaningful differentiation with proprietary data requires innovation that is often beyond the core competency of the traditional platform. Artificial Intelligence is no longer enough. Vast networks of data capture, or DMP integrations are now table-stakes. For real productization, platforms must seek data solutions that move past vague segments, and towards behavior-driven insights beyond the website visit and even beyond the digital experience. The real world presents increasingly sophisticated opportunities for non-PII data capture that given modeling such as Biological Intelligence over the traditionally binary artificial, add new dimensions to predictive advertising that far outshine the obsolescence creeping into programmatic. Programmatic’s legacy of website retargeting, still a dominant technique, will pale in comparison to the potential that mobilization, Internet of Things and other methodologies existing far away from the desktop will enable in not just advertiser performance, but provider profits.

Final Thoughts

If the product direction turns out to be the winning approach for programmatic, it is unlikely that many providers will successfully make the leap. The unique product design and truly scarce data assets will be inherently in short supply. Therefore first-mover advantage comes at a premium. Which platforms will deploy truly unique approaches to data, and in doing so deliver exclusive, proprietary value? The very notion of premium pricing and higher margins requires exclusive differentiators buyers are willing to pay for – which platforms will strive for true uniqueness, delivering data and data methodologies in a closed system that outshines the competition? Only time will tell. But in a stagnating market, on the wrong side of the capital equation, the players in ‘product’ and not platform will deliver profitable growth and fuel the Product Era for programmatic, because an agnostic approach to a revolution rarely finds a following.

photo credit: constant motion via photopin (license)

The Sense of Place: Brain Science Breakthrough and Marketing Revolution


Among the amazing scientific breakthroughs Nobel Prize winners have brought us in the last few years, few hold such promise in the movement towards Biological Intelligence away from all things Artificial.

The 2014 Nobel Prize for Physiology or Medicine was awarded to John O’Keefe, Edvard Moser and May-Britt Moser for the discovery on how the brain considers location and functions as a natural GPS.

O’Keefe first described what he called ‘Place Cells’ back in 1971, but their characteristics and function seemed, well, too good to be true. But with today’s science and forward thinking, and the help of the husband and wife Moser team’s 2005 grid-cell discoveries, he was able to gain worldwide acclaim for his discovery of the truly remarkable capacity of the brain to physically identify ‘place’ with an explanation for the neural mechanisms driving spatial memory.

In their press release about the award, the Nobel Foundation described the discovery as solving for one of humanity’s most complex challenges:

“How does the brain create a map of the space surrounding us and how can we navigate our way through a complex environment?”

While some of the more ‘directionally-challenged’ of us must have this area of the hippocampus less prominently developed, we can all agree that this is a fantastic discovery from the perspective of science, with potentially far-reaching impact to include Alzheimer’s and dementia research – diseases that affect the same area of the brain.

And with discovery of the brain’s remarkable capabilities comes new opportunity to leverage our learning’s for the greater good. Perhaps nowhere greater than in efficient messaging, where today we inefficiently receive 360 interruptive marketing messages per day. With Biological Intelligence defining consumer behavior by ‘place’, and our now improved understanding of its neuro-structure, we can combine the science to further move away from 1’s and 0’s in our pursuit of commerce and communication. It’s the nuanced, biological behavior that predicts human patterns – increasingly driven by time and place. Shailendra Rathore, studying the Subiculum (part of the Hippocampus) in the lab of Dr. Francesca Cacucci, a former PHD student of O’Keefe says “The way people behave in different environments does indeed seem to have influence on spatial representation, I believe that studying what people do in particularly novel and rewarding settings and machine learning these situations from behavioral parameters may reveal interesting information for targeting purposes.“

Biological Intelligence in marketing is built on the idea that collectively, crowds of people move through their days with a singular intelligence – the sum of all intent. With millions of individuals making up the collective intelligence as a ‘system’, we can isolate sub-systems and better understand human behavior by location.

BI first-mover, Locomizer is the only audience platform that is leveraging the identity of place and it’s meaning for the characteristics of frequenting individuals to help brands improve their predictive marketing. With the marketing industry’s dearth of accurate demographic and psychographic data, Locomizer enables a third dimension for brands to better understand their customers. Academia is striving for some of the same learnings. Rathore continues, “Fundamentally we are trying to understand how the brain represents space and memory. Perhaps our conscious experience is explicitly structured spatially. We reconstruct a spatial scene and then move within it when we perform autobiographical recall. Trying to ascertain what is memory and how it is structured in the brain is a key stepping stone towards understanding conscious experience and also building the next generation of Artificial Intelligence.” In the case of Locomizer, AI takes on a ‘BI’ scientific approach.

Biological Intelligence is not only defined by human behavior, but it can be the collective intelligence of systems as seemingly chaotic as, say, ‘cells’. And until now, one could better argue that the daily movement by humans through time and place had equal elements of chaos, driven less by collective intelligence and more by unpredictable causation, aka the randomness of our daily lives.

But with the brilliance of O’Keefe, Moser and Moser, and their discovery that our ability to navigate complex environments is a lot more sophisticated – and therefore intentional – than previously assumed, we can leverage Biological Intelligence for predictive marketing with even more confidence. Whether it’s our navigation through shopping malls – indoor or outdoor – city streets, neighborhoods and suburbs, the chaos driving our patterns and therefore our commerce, is quickly rationalizing. And Biological Intelligence has the formula.

photo credit: Brain diagram via photopin (license)

It’s the Data, Stupid!


As the 2016 American election season starts to heat up, and primary candidates fight for position in their party, it isn’t unusual to hear the oft-remarked James Carville quote “It’s the economy, stupid.”  It came during Carville’s strategizing for Bill Clinton’s successful 1992 campaign against George H. W. Bush.  It may not be the most eloquent quote, but it’s one of the most memorable.  In today’s programmatic ecosystem, vendors are lining up for agencies, brands and, well, other vendors.  They’re lobbying for position just like a political candidate.  Some add this, others add that, but far too many forget what could also be defined so aptly:  “It’s the Data, Stupid!”

We’re now up to 2000+ marketing technology vendors in the industry.  There are E-Mail platforms, Display Platforms, Marketing Automation Platforms – platforms for just about everything.  And while each facet of the landscape has a unique pitch, they all require targeted data to execute.  Content without an audience calls to mind another famous quip:  “If a tree falls in the woods and nobody is there to hear it, does it make a sound?”  With poor data, that’s the question many campaigns have to answer.

Programmatic currently solves the data challenge with complexity.  Where there is poor data, there is optimization.  Where there is no data, there is prediction.  Where there is great data, there is personalization.  This is a good effort.  However, with mixed results, and a lack of demonstrable success in programmatic, many brands are still steering clear of pouring budget into the space.  And many marketers are yet to understand it.  If the complexity, and blurred ROI comes from compensating for poor data, better data should surely solve everything.

So what constitutes better data?  Programmatic essentially has two categories of data.  First, segment data.  This is data typically associated with a demographic.  It’s largely B2C.  Next, is behavioral data.  This is data associated with an action, such as a search query or abandoned shopping cart.  It’s much better but too scarce on which to build a multi-billion dollar industry.  Individual brands can pay a premium for behavioral data and find success, but the industry at large must find another answer.

Quietly building steam is another category of data.  In fact, it’s so smoothly entered out vocabulary that we hardly realize it’s the first data source that’s actually native to programmatic – behavioral and segment data have existed since the beginning of digital marketing.  The new category is geo-data, or even geo-behavioral data, the amassed non-PII information generated through mobile, social, beacons and other local pings. Geo-behavioral data gives programmatic it’s best opportunity yet to solve its challenges.

Most readers will say “geo data is simple.”  “I can use geo.”  But that’s exactly the problem.  Early usage of geo-data has been just as it sounds – data about where someone is or was.  It might try to push a coupon if you’re in the mall, or some similar gimmick.  In a worst use case, it might utilize segment data to determine user demographics by where they are.  Because geo-data is native to programmatic, it needs another dimension to be activated.  It’s not just about where someone is.  It’s about what that place represents, when they are there, who else is there and how they are moving through that space which ultimately solves for ‘why they are there’.  In assessing geo-data, programmatic should realize two concepts:

  1. Each place has it’s own identity, which can change at different times of day
  2. A consumer’s geo-behavioral characteristics are defined not just by the place identity, but by other consumers there at the time. We are a product of who we hang out with.

If these facets are important to the simple, successful processing of geo-data, to augment the poor segment data and scarce behavioral data for programmatic, then why aren’t they more widely used?  Because analyzing this data contradicts a lot of the traditional AI/Machine Learning algorithms on which programmatic was built.  Traditional binary computing struggles to identify such human nuances as passage through time and space with a crowd.  So how can programmatic improve its process to activate this data and make profitable use of it?

Biological Intelligence is the simplest solution to enabling marketing tech to digest, make sense of and scale geo-behavioral data, the missing component in programmatic.  BI, considers human behavior and the power of association and affinity as a native computation.  Biological Intelligence has a major advantage at it’s core:  It understands that systems have intelligence and a collective behavior, even without a brain (no group has one brain).  Therefore, instead of trying too hard to predict one consumer’s behavior in isolation, BI can effectively crowd source behavior, and save a big computational headache.

With programmatic iterating into the future, the complexity and dubious results many marketers are experiencing can be traced back to one thing:  poor data.  However, trying to solve the data conundrum with the two standard data types will prove difficult – especially as programmatic scales to demand even more scarce data.  The only solution is to improve the use of geo-behavioral data, and in doing so leverage new ways of analyzing that data.   In an industry that uses “Intelligence” all too often, it’s an interesting contrast to see the word ‘stupid’ in a post about data.  But until we realize the potential of new data sources, our technology will be just that.  Because of it’s superior methodology for unlocking our richest data set yet, Biological Intelligence, and the easy interpretation of geo-data should get everyone’s vote as programmatic seeks new ways to drive and prove ROI for a simpler, more marketable future.

photo credit: Sunny Pubs in Google Earth via photopin (license)

Geo Data is the new Yelp


At the end of his early 2014 article in ‘StreetFight’, which focuses on the business of ‘Hyperlocal’,  Steven Jacobs made this point:

Social recommendation products from Facebook and Twitter could threaten Yelp’s position in the market, but the longer-term threat to Yelp comes from a more data-driven approach. As mobile positioning and battery-life improves, and devices can persistently track consumer movements, there’s an opportunity over the next decade for companies to use those data to facilitate discovery similar to Amazon recommendations. Content isn’t going anywhere soon; but data will have its day.

The larger premise was a discussion of ‘Perfect Information’, which in economics is ideal for the educated decision-making of consumers.  In business, it can lead to commodization, as brands struggle to differentiate and often mimick the services of competitors.  Take, for example, the airline industry.  Near-perfect information about the features customers cared about, led to lots of price-comparing platforms, and the eventual commodiziation of the travel offering.  But in the local retail market, better information will drive more commerce and enable upstarts to thrive, which is a step in the right direction.

Jacobs’ point really speaks to an evolution in information the mobile web will enable, as non-social data begins to match social data in terms of relevancy, accuracy and supply.  Social and publisher content has been essential in filling the gaps that business and location data leave because without updated data points in local such as store hours, offerings, sentiment, location and more, social content is a useful substitute.

Before the internet, and throughout the early internet, the Yellow Pages were the resource for such data.  Whether in the book, on their website, or via phone, this old-fashioned approach sought to provide consumers with the information they needed to make decisions.  It was updated once or twice a year, and unless they bought nice advertisement, did little to drive demand.  As Jacobs pointed out, this culture of information led consumers to trust primarily large brands, such as “Hilton” for their known quality – the information to advocate for smaller brands just didn’t exist.

Enter social.  Social content and sites such as Yelp, Trip Advisor and many more empowered people to curate the information, data and content of businesses in their locale and those they visit.  With a social sense of community, and hopefully enough iterations, these platforms could produce a reliable proxy for verified data, on all things important to retail.  And not only does this content illuminate where the business is, when they are open and what’s on the menu, but it goes a step further, to depict the kind of traffic a business or commercial center was seeing.  With high review volume a reader can be more confident that the business in question is worthwhile.

Non-social, Geo-Behavioral data has the ability to do the same thing, at scale, with less reliance on the nuances of social authorship.  By tracking non-PII consumer movements locally, including time spent, and affinity with other known data points like sporting events, concerts and more, geo-profiling at scale will fill our local information sources with automatically curated data – which in many cases will prove to be more reliable.

The reliability challenge with social content is that it often requires a polarizing subject to receive volume.  Therefore, people must be passionate about it – whether or not it’s role in the community is truly superior.  For example, a fun Frozen Yogurt shop will receive more content than a reliable breakfast joint that serves an older demographic and certainly more than a local oil-change garage which has served the community for decades.  With improved data automation through geo-locational aggregation, non-polarizing yet high quality businesses will regain equal footing with the trendy places.  Social content does a great job of communicating what we love and what we hate, but a poor job of communicating what we need.

With more iterations, driven by the world-wide proliferation of mobile devices, more instances of mobile ‘check-in’ and interaction and better data capture from both structured and unstructured sources, real-life ‘data’ will start to communicate as much about local businesses as real-life sentiment through social.  And through automation, local behavioral data will paint a very accurate picture of both essential business information and the more nuanced as algorithmic modeling depicts interesting facts about a place such as ‘Best for Sporting Events’ driven by unbiased, non-polarized real-life data.   Many small businesses post the sticker “Think Global, Act Local” on their windows.  The data ecosystem is enabling just that as local actions will drive more data and insights globally for a new level of near-perfect information for retail.

photo credit: Tokyo : 9 Mar 2012 via photopin (license)

An Artificial Intelligence Alternative for Replicating Emotional Intelligence

Biological Intelligence as Artificial Intelligence AlternativePaul Frampton’s argument for improved Emotional Intelligence should be taken seriously as marketing requires an increasingly creative approach in increasingly real-time social channels. Yet in our industry’s pursuit of all things scalable, we try to recreate instinctive human ‘EI’ with the non-human, automated and hopefully scalable AI. Unfortunately, despite sophisticated algorithms and plenty of investment, these AI based algorithmic approaches cannot yet command a fraction of the emotion garnered by creative marketing.   Many will argue that it will only be a matter of time before we see a breakthrough and we should continue to iterate forward. However, we are actually solving this challenge incorrectly. Artificial Intelligence will never catch up because while binary computing is effective at crunching numbers, it is fundamentally poor at predicting human behavior. It’s time to look elsewhere for emotional intuition and luckily an alternative approach – Biological Intelligence – offers a solution that is starting to gain major traction.


The biological approach to solving such challenges has been overlooked for years because many presumed that Biology is mainly a ‘descriptive’ discipline rather than ‘predictive’. However, the evolutionary, genetic, neural networks and swarm algorithms are just a few examples where the workings of life were perfectly formalized in the form of a theory and equations. Now we start to witness how quantitative and theoretical Biology could be used to solve current and future problems in all aspects of our lives.


In understanding how this can be applied in marketing to define the behavior of consumers, Biological Intelligence begins with an important learning from nature. Systems, from groups of cells to crowds of people need not have a single, collective (or any) brain to be intelligent. The collective decision making of the group is not conducted in the way we make decisions as individuals but rather adaptively and collectively, driven by internal and external stimuli to which systems react for survival. The first step to recognizing the superiority of BI is to stop trying to replicate the human brain and observe intelligent systems.


And this is a much sought after pursuit in traditional marketing – with or without an algorithmic approach. Joseph Vita DeLuca, VP Marketing and Communications at Yieldr, aptly described the challenges and opportunities of better understanding separate biological systems in the marketing sense as ‘hypersegmentation’.   The basic principle is that brands can communicate best to audiences who are properly segmented, therefore having enough commonalities for relevant messaging. That is harder than ever today however as segments must adapt for rapid change. DeLuca states Part of incorporating a hypersegmentation strategy is creating an automated process that habitually incorporates new data points while refreshing older ones, in order to optimize messaging through every point of the customer journey.” It wasn’t mentioned in the article, but one way to achieve that automated process is with Biological Intelligence.


Such an approach is increasingly important given Artificial Intelligence’s limitations in predicting ‘irrational’ human behavior. Nick Seneca Jankel, creator of ‘Breakthrough Biodynamics’ defines the issue quite simply. Artificial Intelligence assumes that people will behave in predictable ways, whereas true ‘breakthroughs’ are unpredictable. He points out that Stuart Kaufmann defines this as ‘Partial Lawlessness’ which we can take to mean ‘unpredictable’ or ‘irrational’ decision making within otherwise logical patterns. Consumer behavior is ultimately very unpredictable and trying to solve it with AI based on rationality will deliver the same frustration as trying to accurately predict a football team’s outcome based on common sense – actual outcomes don’t play out that way.


While effective marketing through intelligent decision-making is a difficult task, pursuing it with Biological Intelligence is a move to the simple from the complex. The simple and natural behavior of systems is a refreshing panacea for the quirks of Artificial Intelligence. In this case, a natural approach is better and consumer behavior patterns can ultimately be defined by simple and universal rules. Switching to Biological Intelligence from AI represents a simple opportunity for a paradigm shift in the predictive and analytical results of marketing technology.


About Alexei:

Locomizer Founder Alexei A. Poliakov holds PHD, Biology from Moscow State University. He has dedicated more than ten years of scientific research on spatial behavior in live systems.

photo credit: Terminal via photopin (license)