« Back to Posts

video television

The Revisionists: Zeev Neumeier, Making TVs Smarter Than Ever


The Revisionists is a series that looks at the minds that shape the world we see: TV veterans, agency leaders, innovators in content, analytics and more.

Today, we meet Zeev Neumeier, who’s making televisions smarter for brands, programmers and consumers. 

Zeev Neumeier, SVP Technology for Inscape Vizio, is the kind of guy who won’t back away from a challenge just because, “It’s always been done that way.” He’s blessed with a touch of intentional indifference toward his odds of success in overturning such habits and approaches. Consequently, he tends to tackle challenges where he has no business winning. Sometimes, he even manages not to lose.

Inscape Vizio ACR analyticsZeev found his way into the TV measurement space almost a decade ago. Back then, representatives from a large OEM walked into his office, wanting to hire him to build widgets for a then-newfangled technology called Smart TVs.

After listening politely, Neumeier said, “That’s nice. But who cares about widgets on a TV? Let’s make a proper onscreen interactive experience synched to TV shows. And by the way, the data from that experience? It would be even more valuable than the experience itself.”

The OEM’s reps said, “You can’t do that. And besides TV measurement is done with surveys and panels.”

When Zeev told them that was a dumb way to do things, the OEM’s reps said, “If you’re so smart, fix it.” And thus, Inscape, the smart TV-measurement company, was born.

Inscape Services started as Cognitive Networks, which Neumeier founded and later sold to Vizio, which rebranded it.

Zeev now serves as Inscape’s SVP of Technology, continuing to iterate on new ways to turn Smart TVs into Smarter Ones. Zeev studied mathematics and computer science at Yeshiva University before receiving his MBA from New York University. Zeev developed one of the first automatic-content-recognition systems (ACR) and holds 10 patents.

What inspires you?

Problems inspire me. Or more accurately, a lack of decent answers to problems. If something has a good practical answer, I am happy to use that answer and move on. If something (suspiciously) does not have an answer (or, more frequently, has a terrible answer), then it’s worth my time, sweat and toil to make it better.

The world is imperfect. Big deal. If it was just imperfect, it would be obvious how to fix it. But alas the world is imperfect in imperfect ways where the “fix” is elusive and the distance between what looks good on paper and what actually works in reality is often full of these warm, squishy, walking things we call “people.”

So what inspires me? When an answer works, when I feel that moment where a solution clicks, when the germ of an idea combined with endless iterations and countless suggestions, is then painstakingly reduced to practicality. When that idea finally kicks into life, is used, and then gains wider acceptance, that’s what inspires me.

What are you doing now?

Inscape is an industry leader in advertising solutions through automatic content recognition. We capture live viewing data from content sources across millions of Smart TVs and devices. Our comprehensive metrics deliver highly accurate information about cross-platform viewing behavior.

We provide advertisers and content producers with important insights that help them develop a deeper understanding of audiences, make more intelligent ad-buying decisions and better prepare for market changes.

What’s the biggest change you’re focused on?

Everything changes all the time, all around us. It’s easy to forget that the fundamentals remain the same. Even the idea that change is not new is not new, as Ecclesiastes 1:9 will gladly attest.

The TV advertising industry is trying to solve a problem eloquently postulated in the 19th Century by (department-store magnate) John Wanamaker: “Half the money I spend on advertising is wasted; the trouble is, I don’t know which half.”

Mediums have changed. Technology has advanced. Newer, better, faster, larger datasets are available. But the fundamentals are the same. Defining the correct audience, reaching it, and measuring the impact is still the key. What does this mean for TV? Glad you asked…

Where is the puck going in TV?

Audience-based buying/selling of TV inventory is the next logical step. It is not a coincidence that the Inscape dataset is ideal for this use case. That, in and of itself, is not exactly a revolutionary statement. Where I do think the most interesting change is coming is in the addition of big-data thinking regarding audiences.

As an example, say you’re trying to sell pink cat shampoo that smells like elderberries. Who’s your audience? Your agency will come up with a model. After a lot of thinking, maybe they’ll say you should buy women 18-34. But that’s a terrible way to do things. In fact, the only reason this is even remotely an acceptable way to buy media is because, “It’s always been done like that.”

We can do better. Now say your ad agency is a really cutting-edge bunch. They’re using an ACR dataset like Inscape’s. Maybe they can construct a “cat lover” audience from an online cookie pool and map it to TV viewing. Now you can see where and how to buy so you can reach your audience. That’s better. The definition of that audience is still arbitrary, but it’s a whole lot better than just buying a demo. It’s a step in the right direction.

Now lets take another step beyond that. Maybe I have decent first-party data on who buys my pink elderberry feline shampoo. Maybe I sell it online. Or maybe I’m a store with a loyalty-rewards program (and if I don’t have one, maybe it’s time to work on that too). Regardless, now I can map exposure to actual results, using an ACR dataset such as Inscape’s.

And here’s where the magic of big data comes in. As a math professor once remarked to me, “Even a rather small number times infinity gets kinda big.” Given a large enough dataset, you can find even the most nuanced of relationships. And ACR datasets are really big, much bigger than TV is used to managing, much less leveraging successfully.

Fundamentally, you don’t care that “cat lovers” buy your shampoo. You care that if you reach people in Audience X with Frequency N, they buy the shampoo. What better way to define Audience X than with the same dataset where you originally found them? From here, it becomes a hill-climbing exercise to refine the where and when that will maximize your outcomes, in this case, creating more cats that smell like elderberries.

Long term, we know that the industry will eventually go to an all-targeted, all-the-time world. Here’s another non-controversial statement: Sometime between now and 2,200 years from now, there will be no more linear TV. All ads will be targeted.

If you have a more granular prediction for when that switch happens, I might be open to a friendly wager. Regardless, I propose that the fundamentals will not change. Marketers in this all-targeted, all-the-time future will still be trying to measure and squeeze out the “Half the money” that is wasted. They will still need a nice, large dataset if they want to do that.

What’s the hardest part for the industry?

“It’s always been done like that.” The TV ad world has two kinds of people: Those who use Nielsen even though they know it’s wrong, and those who know Nielsen is wrong but still use it.

I’ve lost count of the times we walk into meetings and the customer spends the first 30 minutes lambasting Nielsen: “Nielsen is terrible.” “They’re undercounting our rating” (Curious how Nielsen’s always, only, undercounting). “Their sample is too small.” “They’re missing our C3 and C7 viewing,” and on and on and on.

Then they ask, “Can your data duplicate the Nielsen number? To how many decimal points?

In Nielsen’s defense, they are a currency. They are not supposed to be correct. They are supposed to be consistent. Being consistently wrong in consistent ways is perfectly fine for their business model.

Nielsen, of course, knows this and is therefore more concerned with defending their castle than they are with moving the needle for advertisers. Even when they do supposedly “progressive” things such as add set-top box data, it’s just there to fix their panel data and bolster the old currency measurement. But I digress.

An excessive reliance by the industry on, “It’s always been done like that,” boils my blood. But it’s the nature of the $70 billion beast. The industry is being dragged kicking and screaming into the future. One step at a time.

What the $%#@?

What can the Bible, Psalm 23 to be specific, teach us about the TV industry?

The crowd-favorite line from “The Lord is my shepherd” psalm is, “Yea, though I walk through the Valley of the Shadow of Death, I will fear no evil: for you are with me.” It’s quite possibly the most recognizable line in Western Civilization. How much ink has been spilled divining the beautiful poetic mystery of Death’s shadow and the hidden meaning of the valley that it apparently inhabits.

Unfortunately the line in the psalm is also wrong.

The original Hebrew word here is tzalmavet, which sure seems like a contraction between the words tzal (shadow) and mavet (death), hence “shadow of death.” But it’s probably not.

The word is used about a dozen times in the Bible and, from context, it clearly means “very dark” or “pitch black.” It’s not clear if the origin of the word involves a death reference at all (as in as dark as a grave or something) or if it’s just a coincidence.

Roots in Hebrew are usually three letters and there are only so many three-letter combinations. My personal favorite explanation is that this mavet is related to the Egyptian goddess Mut and has nothing to do with death at all. Mut might also be responsible for multiple other names in the Bible with an inexplicable “death” postfix.

Regardless, this word has been tripping people up for a long time. Throughout the Middle Ages, various scholars complained that people routinely made this translation mistake. More than 500 years before English poets immortalized the incorrect yet iconic line, Dunash ben Labrat (920-990) was pointing out the mistake, in a translation that was known in both Europe and the Middle East.

Many newer translations have now switched over to “dark valley” as everybody who looks into this agrees this is a better translation and it also makes more sense.

So why is the mistake so prevalent? Simple. It’s very, very old. The Septuagint, translated sometime in the Third Century B.C.E., makes this mistake. Not surprisingly, the Latin Vulgate (Fourth Century C.E.) translation followed suit, as does the King James version and, well, you get the story. Or in other words, the mistake sticks around because, “It’s always been done like that.”

Hopefully, it will take the TV industry less then 2,200 years to come to its senses.

Follow Zeev on LinkedIn.