Subscribe   |   @sriramk   |   About

Thoughts on BAT and ads


The Basic Attention Token (BAT) caused quite the sensation a few weeks ago by raising $35m in under 30 seconds. Depending on your POV, this is either the second coming of Adam Smith or a tulip-dotcom-subprime-bubble monster. For me, it represents an intriguing attempt to use the blockchain to solve some pressing problems in the media and ads landscape. It’s a little mystifying why BAT hasn’t gotten any attention in the ads world or the adtech press. It doesn’t help that most of the paper is either math or cryptoeconomics - not exactly common discussion topics over rosé at Cannes.

Disclaimer: I’ve spent the last few years working on large ad platforms - including the ones the BAT team specifically mentioned - and hold substantial stock in those companies. That probably makes me biased/conflicted (responsible?) in some way. Emptors, consider yourself caveated.

The BAT thesis

The BAT thesis is simple - the digital ads landscape in 2017 is broken. I highly recommend reading their paper (their roadmap and some reddit threads on the topic are also fun reads). In essence, they want to go after three things.

It’s hard to argue with any of these. You don’t have to take it from them or me on faith - all of us have hit the unskippable mobile ad that takes over their screen. You can just feel that things are busted.

The BAT Fix.

BAT’s solution for this is both audacious and simple in design. Instead of letting the publisher measure attention, they propose letting the browser do it as a neutral party (with their Brave browser being the canonical implementation but in the future could be a plugin/extension). And then - here’s the fun part - converting that attention into a *token *with monetary value which can then be transacted directly between the advertiser and the publisher. It’s as if the browser has access to the advertiser’s bank account and sends some money over to publishers everytime it verifies an ad is seen.

By replacing multiple pieces of the typical adtech stack, BAT removes the need for verification and tracking. The browser ensures that an advertiser’s ad is seen by a human being. It also helps make the publisher make more money - fewer middlemen, fewer rakes to pay ( I just love saying “rake” instead of “fee” or “commission” - makes me feel all Bill Gurley-esque). Fewer middlemen also means better user privacy as, simply put, less companies have access to the user’s data. BAT then goes further and rewards the user for their attention (advertisers can send BAT to users) and tries to fix annoying ads by showing relevant ads based on the content on screen (more on this in a bit).

This is a noble (brave?) attempt but I’m afraid they have some very large challenges.

#1 The large platforms work well and at scale.

From the BAT paper

“This “perfect storm” for publishers has only gotten worse over the last few years as Google and Facebook have taken more and more share of advertising revenues. Together they claim over 73% of online digital ad revenue, and an astounding 99% of all growth from 2015 to 2016 in US total online ad budget”

Did you catch that? That’s the BAT team talking of how small the market they’re going after is. Most digital ad revenue and almost all growth is on Google and FB which is a problem for BAT as their ads are native - inline with search results, a part of Youtube videos or a part of your feed. The entire BAT paper’s focus on intrusive ads and tracking is focused on a space that represents less than 1% of growth.

The publishing industry does have concerns about the growth of Facebook’s IA and Google’s AMP and the sustainability of a business model that relies on them as opposed to the open web. But publisher control is a different topic altogether from fraud/privacy and the paper conflates them into one.

To put it differently, there is already a mass adopted fix for two of the problems BAT is trying to solve - the large internet platforms are it. They have scale and advertisers are already directing most of their spend to them.

#2 Brave needs mass adoption

The second structural weakness for BAT is that its success it tied to Brave’s adoption. It is hard structurally for any new browser not tied to a large company to gain substantial marketshare in 2017. Even if Brave gains significant adoption, it still tackles the small part of online spend mentioned in #1 (everyone else being inside apps or natively rendered) - smaller piece of what was a very small piece to start with.

This isn’t impossible - Brave looks well built (Eich was one of my teen coding idols for his work on Javascript and VMs) and I’m actually running it now as I write this. But getting large share in 2017 seems daunting.

[Edit - added below after Twitter discussion]

One argument from BAT is that the user growth pool incentivizes early users to download and install it. I remain unconvinced that it could happen at a scale (given #1) where it will impact advertiser budgets or behavior.

#3 Content-adjacency doesn’t lead to the best ads.

Where the BAT paper is really weak is, is when it comes to targeting. The BAT roadmap has them building targeting based on content. From my experience, relevant advertising that isn’t intrusive comes from understanding the user - not from content adjacency. You are a lot more than the piece you are reading now or the video you are watching. BAT by limiting itself to content similarity for ads is capping how effective/interesting those ads can be. This, at a macro level, is the same issue Apple faces with its ML efforts by limiting how much data can be used in the cloud.

One interesting idea I haven’t seen explored is whether user personalization can be done by the browser. Can the browser store the user’s interest profile and sift through ads and rank them locally? Given the scale involved (ranking millions of ads in milliseconds to find the right one), it seems daunting to do this locally but this seems an idea that deserves further thought.

#4 Paying users for watching ads misaligns incentives.

[Edit: I added this after getting a bunch of questions on it]

One argument BAT makes is that by paying users to watch ads, you get around ad fatigue and structural issues which leads to users not paying attention to ads. I’m quite skeptical of this approach as it misaligns advertiser incentives and user incentives. Advertisers need outcomes - users to do things of their own free will (buy a ticket to a movie, buy that HDMI cable) while users will be incentivized to just watch an ad to earn monetary value.

There is something magical about organically watching an ad and deciding to buy/try something that just doesn’t happen when you bribe someone to do it. Multiple past efforts on this have failed. An easy example is the slew of “offer-wall” incentivized app installs that were the rage a few years ago. Almost always, they yielded low quality users who weren’t interested in actually using the app and churned very quickly.

I could be wrong but this model has a number of historical cases against it working. attention even the right thing to measure?

The above might seem negative on BAT which isn’t my intention - I’m genuinely intrigued by the themes and have utmost respect for the team. Despite all the above, they might still get real adoption for Brave and succeed. In specific, they’re onto one key need that a lot of advertisers would agree with - a common way to measure attention across all the mediums users consume online. A open measure to quantify this by a neutral party will be popular with several advertisers.

But, is that what we should be measuring? Does any of this even matter?

The word advertise comes from the Latin advertere which means to turn to. Advertising, literally put, is about turning people towards your products and services. Getting their attention is a means to the goal but not the end goal in itself. The end goal is for them to use your product - book your airline, buy your clothes, order a ride, buy your phone, etc. One argument against BAT is that standardizing on a “view” is like standardizing on human attention across TV and radio - that finding a common attention metric across disparate mediums like search, a feed environment, a video site, etc is pointless.

We often measure attention in advertising because it is easier to do so than measuring the outcome of those ads. But that is just a proxy - often a bad one - for what is the real question all advertisers are trying to answer “Did my ad spend get someone to use my product?”. I believe the future lies not in measuring attention but in quantifying results.

At the end of the day, you need someone to pull out their wallet and buy your stuff. Otherwise, what’s the point of it all?