Thursday, June 28, 2012

Your personal data is not worth anywhere near what you think it's worth

I see a lot of Root Markets-like businesses. Companies creating a way for people to own their own data and profit from it rather than letting someone else profit from it. The idea is appealing: other people are selling your data, it's your data, why shouldn't you sell it yourself?

But most of the people I talk to don't have a good answer to the basic business question: can you sell your product or service for more than it costs you to buy or make it? In this case, can you sell personal data for more than it costs to garner it?

Well, can you?

The IAB says that in 2011 there was $31.74 billion in US interactive ad spend [pdf]. There were 245.2 million internet users in the US in 2011 according to, using data from Nielsen and the ITU. That works out to slightly less than $130 in ad spend per internet user per year in the US.

Here is a breakdown of this per capita number, by channel, and a guess as to how much is potentially available for third party data sellers:

$ per Addressable
Channel User Market
Search 47% $60.84 $0.00
Display / Banner 22% $28.48 $7.12
Classifieds 8% $10.36 $0.00
Digital Video 6% $7.77 $1.55
Lead Generation 5% $6.47 $3.24
Mobile 5% $6.47 $1.29
Rich Media 4% $5.18 $1.04
Sponsorship 4% $5.18 $0.00
Email 1%   $1.29   $0.97
Total $129.45 $15.21

The $130 needs to pay for several different functions. The $28 for display, for instance, pays for account management, creative, media planning, targeting, media buying, ad serving, analytics, verification, and--not least--the actual inventory the ad is placed in. I'm guessing that the maximum amount available to a company selling data to target display ads is 25% of the ad revenue*. The opportunity to use data to optimize lead gen is potentially larger, while the opportunity in sponsorship, classifieds and search is pretty much nil**.

If this is right, and given the fuzziness of the IAB numbers, it means that there is maybe $1.00 to $1.50 per person's data per month available to data sellers.

But keep in mind that Google does not need your data. Nor does Facebook. They are a large part of the market. Your data is competing with everyone else's data--first, second, and third-party data--for this $1 per month. And some of the data you are competing with is so closely tied to the awareness generating process that it can't be pried away and placed in a 'wallet' somewhere.

Take context. The context of an ad can account for somewhere between 50% and 90% of its effectiveness. Context correlates to demographics, purchase intent, state of mind, and behavior. If you are looking at a review of the new Mac Book Pro I don't need any personal information to make an educated guess that you are in the market for a new computer. I can confidently put a computer ad next to that article without any other data, and the only way someone else can intermediate my guess is by blocking the content or ad entirely. Same argument different data for Facebook, and for much mobile usage.

This means that of the $1 per month much less is actually available to you as a collector of the data.

The original Root business model was to allow users to own their data and rent it out to people who wanted to market to them. The problem: users think their data is worth far more than $1 per month. But $1 per month is all that is available, on average. To a single company, it's maybe $0.10 at best. And then there has to be a commission paid to the new intermediary--the Root-like company. The user ends up with maybe a dollar a year. Nobody cares about a dollar a year. There's no business model. I could even imagine a world where each user was worth $0.20 a month, but that price is still nowhere near where it has to be to have users take it seriously.

There is a business model for businesses that gather data very efficiently. There are several pretty large companies that do this. But they have figured out a way to gather the data for much less than $0.10 per person and to collect data on hundreds of millions of people. The Root model simply costs more per person than the data is worth.

I spent several years of my life trying to build a business that lets people take control of their own data while still leaving a way for marketers to find them. I believe in privacy. And I believe that marketers finding customers is key to economic efficiency. I would love to see someone square this circle, but the Root model is not the way to do it.

* This takes into account the fact that I think the IAB/PwC revenue number is the amount paid to publishers, not the amount spent by marketers. The amount spent by marketers may be 50% to 100% more than that paid to publishers on average. Hard to know. This is an important point though: marketing is much, much more than advertising. The amount that companies spend on marketing in total is far higher than the amount that publishers make from selling ads. There are companies selling data that sell into this marketing market that are worth billions, they are not the focus of this post.
** The best businesses are the ones where everyone else thinks you're wrong. My saying there's no opportunity means that if you have a way to use data to optimize these channels, you may have an opportunity that no one else has seen. I like those.


Anonymous said...

I've always thought the way to tackle this model is not through advertising for the exact reason you outline above.

Instead, it's to focus on individual data matchmaking for e-commerce and recommendations. Allow increased personalization for new users to a given site and solve the cold-start problem for e-commerce retail.

So Banana Republic, say, could offer you 75% off your first purchase if you give them your Amazon purchase history (or, say, your Google search history). They can ideally use that history to generate high-quality recommendations when you visit the site and drive increased uplift as opposed to your visiting the site cold with no data about you. Everybody (well, besides Amazon) wins.

It's an extremely hard market to seed but I've always wondered if that would work better.

Jerry Neumann said...

That actually sounds like a pretty good idea. Think you'd need to see how receptive the client base is--it sounds like something Hunch would have tried. You need to figure out why there is no solution there yet.

Marc Rossen said...

Really interesting post Jerry. I work in the Adtech space on the analytics side of the house and couldn't agree more with the economics you outlined.

My thinking has always been creating a way for consumers to let business know what they are interested in/in-market for and allow the business to "bid" on the consumers business.

Meaning you say on a platform "hey I am in-market a new sofa, this is what I am interested in (quality features, etc)" and then let business sell you on price/features. In essence flipping the business model...

Interested to hear your thoughts.


Matt @matthewphogan said...

I think this is a circle that can be squared. The idea of each user controlling their own data on an individual level, as pointed out, is not economically viable, because a single user's data is just not worth very much. When the data of many (not sure at what tipping point) users is aggregated for analytical and pattern recognition purposes, then a robust lead-gen platform can be developed. The users, by consenting to the platform, are basically saying that they want a better shopping experience based upon their data, and they are willing to hand over their data to prove it.
To take the Banana Republic example, posted by anonymous, to another level; if 10k users allow merchant access to their credit card history (and other pertinent data) through an intermediating platform, the intermediating platform can whittle down 500 users who've shopped at J Crew in the last 6 months and spent on average $1k across 2 visits. Banana, upset that J Crew is eating their lunch, can entice those 500 users to come shop at banana (brick and mortar or online) with in-store cash offers of $100 (effectively a 20% discount on a per visit basis).
The real issue is getting enough users to where this platform generates enough data to be relevant. Hard to entice users if you don't have merchants lined up. Hard to entice merchants to sign up without users. Classic chicken and egg conundrum.
That said, with extremely granulated data being used for lead-gen, a merchant funded in-store cash offer platform makes the economics work for the data aggregator.
I'm currently building a startup/platform to address this very problem. For anybody following this blog, hit me on Twitter (handle below, or above) if you want to discuss further.

davey said...

Why hasn't a company like facebook, Verizon or godaddy built a platform for personal domain management to enable this?

davey said...

Did Root Markets create personal domains ? What did it actually do?

Aswath said...

How about running a lottery on the earnings? Meaning instead of each user getting an insignificant $1, a randomly selected subset of them get a significant amount.

C.P.Spencer said...

Context is everything, its the inspirational moment marketers are missing.

Brett said...

Similar to Klout perks, no?

Nicole Hamilton said...

Great post, and very timely. It is important that consumers understand how their data is being used, and then give them the ability to opt in (and out) to offers that benefit them. I agree that privacy is important (particularly in our business, personal finance). Many consumers are skeptical about where their data is going, due to their actual experience with deceptive lead gen practices, and/or lack of clarity with internet policies. If they understand, and usage is clear, it opens the door to better opt in offers and services.