The Trade Desk Enters The OS Wars, The Personalization Conundrum
1. The Trade Desk Enters The OS Wars
You have no doubt seen the stories about how The Trade Desk, the mega DSP, is hard at work building its own TV operating system, a story the company itself denies, though numerous people I have spoken to assure me that it’s true.
You have also no doubt thought to yourself “W the actual F?” as the plan seemingly makes no sense.
If you were paying closer attention, you’d have also seen a further report—via unser Junge Janko Roettgers—that they are plotting this operating system in conjunction with Sonos.
You know, the full-home speaker system that usually comes with the adjective “high-end” in front of it.
So even more W the actual F.
And while I’m pretty much at a loss for a rational reason for this, I do have a theory on an irrational one.
Allow me to explain.
Why It Matters
Many companies who come out of the digital space tend to forget that consumers have agency.
And by this I mean they forget that consumers are not actually compelled to do anything you want them to do, and that oftentimes they won’t.
It’s partly why Facebook’s video efforts have failed so miserably over the years.
They assumed, for example, that people would watch live video just because it was on Facebook and they were promoting it. They did not even consider the notion that people would think their live video sucked and just not bother to click on it.
So there’s that.
Now back to The Trade Desk.
Having a DSP own the operating system makes a lot of sense on paper, at least from an advertising perspective. There’s a streamlining of processes, better ad measurement and better overall data around targeting and identity.
And that’s often a valid enough proposition in digital where there really isn’t a whole lot of choice (Android or iOS? Google or some random search engine?).
But TV?
Not the case.
Consumers buy TVs for a variety of reasons. None of which have anything to do with advertising.
And that’s the rub—consumers have to buy a TV an operating system runs on in order for that operating system to actually have, you know, customers.
Consumers want very different things than advertisers—better search and discovery, better picture quality, an easy way to rearrange their apps.
Oh, and low prices.
Did I mention low prices?
Because that’s a huge part of the continued success of the smart TV industry, the fact that you can easily buy a 55 inch TV for well under $500. Or even get one for free, from Telly.
How do all those OEMs do it?
One key way is that they all have their own FAST services, which create a second revenue stream for them which then allows them to keep the price of the TVs low—the TV hardware business is a very low margin business and success depends on volume not mark-up.
So there’s that too.
Then there’s the market itself.
Which, in the US, is pretty tightly sewn up these days.
Roku owns a goodly share of the market, via licensing deals, dongles and its own branded TVs. (What percentage is hard to say with any high degree of accuracy—the numbers are all over the place depending on your source—but safe to say they are likely #`1 or #2 in the US market.)
Then there’s Hisense, LG, Samsung, TCL and VIZIO, to list the other key players in alphabetical order.
Amazon and Google are trying, and over in dongle-land the Fire TV stick has decent market share, Apple TV doesn’t, and the new Chromecast has—well, it has good reviews.
But that’s about it—there’s not much room for new players and the average replacement cycle on a new TV is about six or seven years.
Now, as we noted in our free Special Report on the TV OS Wars earlier this year, the rest of the world is where the action is. It is estimated that 40 percent of the TVs sold do not have a dedicated OS.
Most all of those TVs are made by local manufacturers whose market is just a single country. So the biggest independent TV manufacturer in Indonesia. Which could be promising for The Trade Desk and all that, only it’s not like those companies don’t have options.
First off, Google (Android) and Amazon are only too happy to sell them an OS. If they’d rather not go with a big tech company, Samsung, LG, Roku and VIZIO (among others) have ways to license their OS, too.
And then there’s a whole other category of independent white label operating systems like Titan OS and WhaleOS who work with those local OEMs, providing them with FAST services, data, and the economies of scale that come from working globally across multiple vendors.
So there’s that too.
Finally, there’s the story of Roku and their struggles to establish a beachhead in Europe. As in, if this were D-Day, they’d still be stuck on Omaha Beach.
The short version of the long story is they’ve gotten trapped in a chicken-and-egg conundrum. Roku apps are built on a proprietary tech platform and so you need to find people who can code Roku apps.
If you are a European streaming service, it’s probably not worth it to you to spend time and money seeking out people who can build you a Roku app when hardly anyone uses Roku. And if you are Roku, it’s hard to get anyone to buy your devices when you don’t have apps for the services people want to watch.
It’s a vicious cycle, one that even Roku’s alleged offers to build the apps themselves, at their own expense, seem unable to break.
So there’s that as well, which both explains why a US-based OS will have trouble breaking into European and other overseas markets and why The Trade Desk may have decided that building their own Android-based platform was a better idea than buying Roku’s non-Android-based one.
What You Need To Do About It
If you’re The Trade Desk… where to start… first off, if you are actually building a TV OS as everyone I’ve spoken to seems to believe you are, then flat out denying it is not going to help you with anyone in the industry, the broader trade press or even consumers. Maybe you’re not aware of it, but you’re only playing into industry stereotypes of devious Machiavellian ad tech companies.
Speaking of not paying attention, you do have a reputation in the industry of keeping yourself aloof, of not showing up at the usual conferences and all that. Which, at some level, makes sense—it’s not the worst idea to position yourself as being leagues ahead of everyone else. But maybe, just maybe, you are missing out on a lot of valuable intel that way. Decisions made from inside a bubble rarely being good decisions and all that.
Something to think about.
Which brings me to my last point.
Yes, on paper, an operating system controlled by a single DSP that is able to provide greater efficiency across the entire ecosystem while also becoming an indispensable source of data and measurement makes a whole lot of sense.
But have you actually talked to the people you think are going to get on board with this?
Because I will guarantee you that there is no way in hell any of them are going to turn over control of their operating systems to you. They like the control. They see the path to victory, or at least the path to controlling a decent enough percentage of the market, and it doesn’t involve giving up control to an ad tech company. Even if giving you control could, maybe, possibly, ultimately net them more money.
And you know why that is?
It’s because they have agency. Meaning they don’t need to do that thing you want them to do. And that they’ll be all right if they don’t do it.
Better than all right even.
2. The Personalization Conundrum
Yesterday, a friend was reflecting on her experience with the Google suite of apps in her car.
It seems that Google Maps will oftentimes consult her calendar on its own volition and suggest she might need directions to the place she is going to next.
That it does so without her prompting it to do so is at times creepy, though she admits that 9 times out of 10, it’s saving her from having to do something she was about to do anyway. It’s just the fact that it knows so much about her…
This is the general conundrum with all sorts of personalization attempts, something that came up during the panel I moderated at IBC last week.
The issue is how to calibrate that line between “useful” and “creepy”, a line that will be different for just about everyone.
Why It Matters
Personalization in television covers a full range of possibilities, from personalized recommendations to personalized playlists to personalized ads to personalized content on news or sports programming.
It’s quite a range, and one question that comes up over and over is how do you calibrate it to take into account the range of reactions to personalization, which, depending on the person, can range from delight to disgust, and everything in between.
Throw in a lack of consistency—someone who is thrilled to get news that is focused on their particular interests might be horrified to get ads that seem, to them, to be overly targeted.
One solution is to give people the tools to make adjustments themselves, to dial up or dial down the level of personalization they experience.
The problem with that is how do you make those tools simple enough for most people to find them and understand them, let alone effectively use them. (Bear in mind there are a goodly number of people who don’t know you can adjust the brightness on your phone.)
There is an argument to be made (and if you watch the full video of the panel, you can see Nvidia’s Richard Kerris making it) that this is merely a matter of comfort, that we, having grown up without these sorts of features, are programmed to somewhat distrust them or find them creepy, whereas kids who were raised with it will find it completely normal and will confidently adjust the knobs to meet their comfort level.
Which is most likely true, but it begs the question as to what to do in the intervening 50-some-odd years when those of us who did not come of age with that sort of technology are still sentient beings, and how to create value and magic for us without all that backlash.
What You Need To Do About It
If you have any sort of user-facing interface, this balance is going to be something you will need to keep front and center for the foreseeable future.
Be upfront about what you are doing. Give people the option to change the level of personalization they get. Not just very techy people—they’ll figure it out anyway. But people who don’t like to adjust the order of the apps on their TV for fear of somehow breaking it.
Those people.
At the same time, don’t hold back out of fear. There are all sorts of amazing things that can happen when experiences are personalized. Just make sure they’re largely consumer-driven, not advertiser-driven.
Finally, don't discount the magic of serendipity, what I like to call the “90s record store experience” where you accidentally stumble on something you never would have thought you’d like and it turns out you love it.
Figure out how to use the algorithm to recreate those encounters. Not eliminate them.
Thank you.
