Taylor Davidson · The future of big data, ad blocking and tracking

Is big data a big asset or a big liability? That is up to us. Let's consider how we are using data to make better products and ads while dealing with security, privacy, and tracking.
by Taylor Davidson · 07 Aug 2015

I recently did a guest appearance on a podcast talking a bit about me - financial models, Foresight and fundraising - and a lot more about big data and privacy, how apps and operating systems are changing, and machine learning and artificial intelligence. [1]

One of the biggest topics we talked about, starting at 13:51, was about data and how product creators and marketers use data:

We’ve been living in this world where more and more data can be a key competitive advantage for a company. … but perhaps we’re seeing a reaction to the quest for more data, for data that’s hard for us to see how it has a direct impact on the quality of the product or service we receive.

As product creators, we’ve rationalized our collection of data as a valuable part of building a better user experience, based on the premise that the more we know about you, the better product or service we can deliver for you. But we’ve also collected it because data creates the adjacent possibilities of new products, business opportunities and revenue models. Data is an asset, and collecting data - and the opportunities that comes with it - has been a strategic imperative of most of today’s new web businesses.

But collecting data also comes with a responsibility: the responsibility to use it securely, without bias or discrimination, to respect users’ ownership rights, to abide by stated data and privacy policies, and at a deeper level, to use it as a force for value creation - building better products - and not just an inert, unused, potential asset.

And many companies are failing at these basic responsibilities. Just look at what happens every time there is a new hack and more of our sensitive personal information is stolen: an apology, an offer for credit and identity monitoring, a promise to do better, and then the hackers move on to the next company. More data stored means more data to be stolen. Big data is a big target, and maybe, a big liability.

* * *

We’re seeing the glimmers of companies begin to make product decisions that have the idea of “small data” as a key component.

Cotap, an enterprise messaging app, launched Clear on June 30, calling it an “important step in protecting users from security breaches by default”:

Cotap Clear gives businesses the option to set the timing of data storage to 7 days, 14 days, 30 days, 90 days, 180 days, one year or forever. The timing can be set company-wide or be specific to certain groups. To keep all of our users safe, the default for storing messaging data is just 14 days.

… Much of the communication happening through technology at work today feels like a conversation you might have with a coworker in the hallway: quick and conversational. And in the wake of a seemingly-endless stream of high-profile data breaches (LastPass, Target, Home Depot, Sony and others), we think it should be up to businesses to decide if that is the type of information they want to store forever. Cotap Clear lets businesses securely wipe data from our servers and users’ devices after a set period of time, putting them in a position to prevent breaches instead of cleaning up after them.

Less data stored, less data to be stolen. The idea has shown up in other areas and ways: tokenization to reduce the amount of credit card data stored by merchants; email, SMS and app notification methods for user verification replacing password-based authentication; ephemeral messaging for consumer and business use; data models built in the cloud but optimized to be used in low-to-no-bandwidth environments (example, Google Translate using neural nets to do real-time visual translation by optimizing data models for the phone), and more. Apple has been has been clear through their announcements and product decisions that they don’t want to track what you do when you’re using its products, a position that could stand to be prescient. As the number of devices connected to the Internet stands to multiply and the opportunities for Internet of Things data to be tracked, stolen, and misused in ways hidden behind free services and terms of service agreements, the risk to companies like Apple of major security and privacy breaches only continues to grow.

The near-daily stories of new data hacks should give us all a pause, and perhaps a rethink of what data we choose to store and use. We have a responsibility to use the data we collect to build better services, and if it truly doesn’t help us build a better product, we shouldn’t store it. [2]

Data can no longer be an asset that we store indiscriminately. We’re not facing a technical, bandwidth, data storage or management limitation, but the real risk of a cultural shift, just a black swan event away from waking up to a very different perspective.

* * *

Which brings us to advertising, and all the fun stuff the advertising industry has been talking about lately, namely spam, bots, ad viewability, and adblocking. [3]

The ability to use data to target ads to users based on data - starting with usage data, imputed audience data, behavioral data, or observed intent data - has driven the rise of digital advertising and created and supported the third-party adtech ecosystem.

Bots, spam and viewability trouble advertisers because they don’t want to pay for ads that aren’t seen by humans, and some parties in the ecosystem are abusing the ease of launching sites and running ads to skim millions of dollars from advertisers. It’s a problem the industry is working on solving, but it’s not necessarily a problem that users - the people being advertised to - care about.

What users do care about is what ads they see, why they see them, and more recently, figuring out ways to not see them. While the debate over ad blocking often devolves into a debate over the morality and ethics of blocking ads, the fact is that the user adoption of ad blocking technologies is an expression of what people want: and it’s on the rise. Why? One, it’s coming closer to reaching a critical mass of the US online population, with over 40% of people using ad blocking technology in some way, and two, Apple’s coming iOS 9 update includes the ability for users to use third-party “content blockers”, which could lead to developers building third-party ad blockers for mobile. Should this be a big concern for advertisers? Maybe, maybe not. My gut says it’s probably not a huge concern immediately, but it’s a sign of where the tide is heading.

Remember Napster? Remember when the music industry revolted against piracy and sued people who shared and downloaded pirated music? The music industry forgot that piracy was an expression of how people wanted to purchase and listen to music, and after new services emerged to provide new ways for people to pay to listen to music (streaming, ad supported), piracy became far, far less of an issue.

Is every person that reads a page while blocking the ads truly a stolen pageview and lost revenue? No. The web offers an enormous amount of places for one to get content, and if you, as a publisher, were to force someone to accept an ad, it’s fair to say that some people would just go somewhere else instead. But publishers aren’t giving up, and are trying their best to deal with adblocking. One method is to recognize visitors using adblocking and use alert messages, popups or content walls to educate them on how ads supports the production of the content they are reading. Another is to use a content wall, paywall, or register wall to block access to content unless someone disables adblocking, pays, or registers and logins to the site.

But it’s not just ads: people are waking up to the data that is being collected about them. They’re becoming aware of tracking technologies and how data is being created about how we used the web, and then bundled, sold, and used to target ads back to us. We don’t necessarily understand how the data is used, but we’re starting to figure out that slow pageloads and high data usage isn’t from the content we read but the data that’s being collected on us. And while we don’t have a lot of easy ways to rebel against it, becoming aware is the first step.

* * *

What’s the solution to our aversion to ads?

It’s not simply “make better ads”. If people wanted ads, advertisers wouldn’t have to pay for them. It’s not possible to make enough great ads to make people want to be interrupted by them as much as advertisers want us to be interrupted. Making better ads allocates returns to advertising relatively, but won’t raise the fortunes of all advertisers absolutely.

It’s not native advertising, at least in the long run. Many publishers have taken a hard look at what kind of ads they run, and have chosen to invest into building advertisements that look like content, either as sponsored posts, branded content hubs, content campaigns, or “native advertising”, advertisements that built to be native to the experience of the content being provided. These ads run in the feed and are formatted similarly-but-not-exactly like content, and feel more like a unit of content than something a brand paid for.

But it’s naive to think that running the ads in a feed or making them look like content will abate our interest in seeing less ads, “cure” our banner blindness, or kill technology’s ability to figure out what’s an ad or not. There’s too much consumer interest for the innovation in ad blocking to stop; we’ll figure out ways to use technology to figure out how to discern if something is an ad or content. Or we’ll shift how we access content, reading more through Instapaper, Pocket, or the coming Apple News to strip out the formatting and ads on a page and just focus on the content.

Ads delivered programmatically through exchanges often get a bad rap, but it’s not fundamentally because they are bought programmatically or delivered blindly to publishers through networks or exchanges. The programs aren’t the problem.

* * *

In fact, the problem may not be the ads, but the data we collect in the course of delivering ads.

I find it instructing in how Facebook, Snapchat and Pinterest are taking different approaches to creating ways for brands and advertisers to deliver messages, content, and call-to-actions to their userbases. I’ve written extensively about the value of intent data and the opportunities for intent-rich sites to create advertising and marketing products and revenue streams, and while it’s been understandably hard for these consumer sites to launch their advertising businesses, we’re starting to see moves. And they aren’t always towards the type of advertising we’d expect.

Facebook already has an established, highly lucrative advertising business driven by their logged-in users creating first-party data, matched with third-party off-Facebook data, to create a set of advertising products that fit what advertisers want. Facebook’s moves to integrate more data and targeting into Instagram, as they prepare to launch Instagram’s API and eventual display ads business, makes complete sense based on their own history and data-driven past.

Snapchat, however, is taking a different route towards building an advertising business, one that is highly reflective of their own path. Snapchat is famous for disappearing messages, but also for the data it never collects about its users. For their users, who are growing up being skeptical about how their data is being used, it’s an intelligent choice of a monetization model by Snapchat.

[Snapchat CEO Evan] Speigel took a dim view of targeted advertising, adding that “not being creepy” is a top priority for the ephemeral messaging firm. Snapchat’s solution to the “creep factor” is 3V Advertising, a new ad offering that plays vertical video ads on Snapchat’s “Stories,” or aggregations of user photos and videos curated by Snapchat. The ads will leverage some data, like geolocation for targeting, but will largely be curated by Snapchat’s editors.

…Rather than leaning on gobs of data to personalize ad experiences and finely slice targeting, Spiegel insisted that Snapchat is not amassing large data sets on individual users’ personal information. Spiegel said he has an intense dislike of the kind of targeted ads that are becoming more prevalent. (link)

Focusing on content that’s curated by editors, forcing brands and advertisers to invest in creating content specifically for Snapchat, and staying away from heavy user-defined data-driven advertising, is the right choice (for now) for Snapchat. Snapchat’s users use the platform for a specific reason, and staying true to user behavior is important. The real question: is this approach right for other new consumer platforms, or only Snapchat?

Pinterest is taking a different approach, one that is highly data-driven, but monetized via commerce, not advertising. In many ways Pinterest is leading the emergence of “buy buttons” on social media, and has explicitly chosen to focus on building commerce products for brands that drive purchase conversions on Pinterest itself, rather than building advertising products to drive people to retailers’ sites. Pinterest has been clear that their vision for “Sponsored Pins” is for them to be content rather than ads. They’re paying attention to how their users use Pinterest, and creating a monetization path that makes sense for them.

Collecting more data about the user isn’t at the core of Snapchat or Pinterest’s strategy. Immediacy or intent, respectively, is more important. I don’t know if it’s the route for all consumer platforms, but it’s a signal.

* * *

The problem for most publishers is that they don’t have an obvious, unique path that makes sense to them. Selling content to users isn’t going to work at scale: people aren’t going to pay for content (either through subscriptions or individual content purchases; the NY Times is an outlier) in enough volume to support the amount of publishers we have today. [4] Smart publishers understand that, and are developing other ways to engage their readers. Business-focused sites have an easier path since conferences and events are a powerful and lucrative way to provide additional content, community, and connections to their readers, but even non-business sites are using memberships, events, and commerce as ways to provide value to readers beyond content. But without a logged-in user with at least basic time-series usage data, and perhaps interest and demographic data, publishers are always at a disadvantage to the social networks that also play a central role in their social lives.

Direct sold, premium advertising is lucrative, but takes a sales staff that not every publisher can afford. Network and exchange-based advertising is the form of advertising that people are increasingly trying to block. Branded content and native advertising has a long way to play; the history of advertisers integrating brand sponsorships into content dates back to magazines, newspapers, radio, and TV, so it’s obvious that it would end up online. But perhaps never before have people had the ability to separate the delivery, presentation and consumption of the ads from the content.

I don’t know where ads are going, but if you want to figure out the future of adtech, look to where consumer technology is going.

The reason is simple: people define the nature of the ads they see by the platforms they use, because brands follow attention. Consumers “choose” what’s popular based on where they direct their attention, and brands find ways to tap into that attention to market their products and services. Brands are forced to adopt to new platforms and new attention norms because if they aren’t, their messages won’t resonate and their advertisements won’t perform.

That’s why I’d pay a lot of attention to mobile messaging, i.e. texting and messaging apps. What Facebook does with WhatsApp is probably more interesting regarding the future of ads than what they do with Instagram. The rise of text-based concierge services is particularly interesting because it a) leverages a behavior that is almost unbelievably popular and deeply ingrained into digital behavior, especially among the young, and b) we may not see interruptive ads in messaging, but we’ll see some ways for companies to provide services and make offers to people. [5]

Companies have to provide what people want, or perish. People want products and services that provide them more value than they cost. Maybe we care about how it’s made, or where the materials come from, or how the workers are treated. Maybe we care about what they do to try to market to us. Maybe we care about how they use the data they have about us, and how they store it or keep it safe. We have the power to determine what and how companies build for us. In the long run, we define the ads we see, not advertisers; we define what we’re sold. It’s up to us to decide.

* * *

UPDATES: Lots written about adblocking after I wrote this. Alex Kantrowitz explains how ad blocking isn’t really leading to the $21.8 billion loss that many are claiming. Darren Herman explains how ad blocking and tracking aren’t the same thing. Ad blocking popped up on the Today show. Eric Franchi offers on his thoughts on how adblocking could create a flee to “media that matters”. Doc Searls wrote about content blocking as the chemo for ad tech. I agree with Joe Marchese, quoted in AdAge: “We can block the ad blockers, but technology always finds another way.” The turmoil is the shifting balance between what consumers want and what the market is providing to them. We’re hearing the conflict between the incentives behind billions of dollars and thousands of careers. Let’s be clear: ad blocking may change adtech, but advertising, even data-driven advertising, isn’t going away.


  1. Btw, I’m really looking forward to (hopefully) doing more podcasts. Working on improving my radio voice right now. ↩︎

  2. This seems like the right place to talk about what third-party tools I use, data I collect, and why I do it. I use Google Analytics for analytics and Twitter analytics for Twitter link sharing and conversion tracking, both to help me understand how people use my site and what they share from it. I use Waypoints to trigger certain content popups and Ouibounce to power the exit intent email subscription popup. Waypoints stores no data and uses no cookies; Ouibounce uses cookies to ensure that the popup is only shown to users once every 30 days, and that if someone signs up to never show the email subscription popup again. I use UTM Form to track details about the source of email subscription signups, primarily to track how the subscriber found my site and specificaly where on my site the subscriber converted from. I use the data from Ouibounce and UTM Form to inform my decisions on how to provide a site experience that’s balanced between a user’s desire to read content as quickly and easily as possible and my desire to understand and build a relationship with my readers. From time to time I run offers on products and services I provide via Foresight, and while I have run ads in the past from quality ad networks, at the moment I am not running external ads. ↩︎

  3. I used to be VC focused on advertising and marketing tech; while I’m no longer focused purely on adtech, it’s still an area I follow closely. ↩︎

  4. Maybe that’s the point, though. ↩︎

  5. Read What SMS is good for and Futures of Text if you want to understand this space. ↩︎