A model for dealing with ad blockers

First, value your audience. Then think about how to treat them.

An interview on the sidelines of a major media conference this spring led me on an improbable journey.

I had already applied to the journalism-related JSK Fellowships at Stanford to examine how online publishers should address ad blockers — those free pieces of software that block the web’s annoying ads but simultaneously cut off the flow of online ad dollars to writers and editors like me.

The interview forever changed how I think about this problem.

The executive had just gotten off stage from a Q&A about his company’s ad-free service. The $4 per month subscription removes ads from the site’s video stream. Meanwhile, anyone with an ad blocker is asked to turn it off, and if users don’t, they have to wait anyway for the ad segment to finish, while staring at a black screen.

It’s an interesting strategy and not everyone knew then whether it would work.

The question is incredibly important.

See, there are hundreds of millions of Internet users who are blocking ads worldwide, and the number is only growing. This group of rebels remains a giant threat to anyone hoping to make a living online now and in the future through ads. That’s pretty much every journalism outlet on the planet.

Interestingly, he told the audience that ad-free subscriptions were growing, right alongside ad revenue.

I button-holed him afterward.

How did you pick the ad-free price? Is it higher than the average ad revenue generated by each visitor?

No, he said. It’s below the average.

Huh.

Turns out, people who see lots of ads (and are therefore worth more to the service) don’t mind seeing more. They’re not enticed to pay to get rid of them. But certain people who really hate ads will pay, even if they are effectively trading up to do so.

Curious.

In other words, he said, the likelihood that people will take the offer isn’t related to how much they’re worth in ad dollars. You can hire some consultants to run some financial models and pick the price that is forecast to get you the most revenue overall. Some people will trade down, some will trade up, but when all is said and done, you increase the size of the revenue pie.

At this point, my mind is racing.

Follow-up requests for information didn’t get me anywhere. As any reporter will tell you, talking to someone in person elicits a lot more information than trying to go through email. But the cat was already out of the bag.

I cast around the Internet for what kind of modeling he could be talking about. And then I found it.

Monte Carlo.

It’s the name of a great European casino destination. It’s also the catch-all phrase for a form of modeling that essentially rolls the dice on things that you don’t know and sees what happens when you do so thousands of times. (Here’s where being a reporter who once covered gambling in Las Vegas comes in handy.)

If what he’s telling me is right, there are two dice that roll independently of each other.

One represents demand for ad-free at a given price, and whether someone takes such an offer. The higher the price, the less chance that someone takes it. So let’s say at $4, someone only subscribes when you roll a six, but if the price is $3, they’ll take it every time you roll a five or six. The other die represents how much the person is worth in ad revenue if they don’t take the offer. Let’s say $1 to $6 a month.

Which price do you choose to maximize revenue?

A bunch of YouTube videos later, some late nights at my computer and I had a working model.

You see, a personal computer is an amazing thing. Today, every person in the world with a copy of Excel on their laptop can create something similar. Excel generates random numbers, allows you to apply them to various formulas, and will happily run thousands of calculations at a keystroke.

I already had a pretty good idea of what the distribution of ad value was over an audience — most people aren’t worth that much in ads. But loyal, frequent visitors are worth a lot. The audience generally follows the 80/20 rule, I was told by another executive, where 80 percent of your revenue comes from 20 percent of your customers.

And I know from Econ 101 that as the price rises, demand falls. Just watch how raising cigarette taxes cuts consumption.

Mixing these two inputs together gave me a model that spit out an optimal price.

But I’m still just a guy. I needed validation.

When I got the fellowship and came to Stanford, I searched for any course description with the words “Monte Carlo” in it.

I found one taught by an adjunct professor by the name of Sam Savage, who had written a book called “The Flaw of Averages” and is executive director of a non-profit company called Probability Management.

Do you think a guy like this could help?

You betcha.

For the last couple months, Prof. Savage and I have been working on the model in bite-sized bits of time in his office, via webcast and over lunches at Forbes cafe (the only animal protein he eats is octopus, but that’s another story).

The professor describes this modeling work as akin to flying an airplane, or riding a bicycle. No matter how many formulas for airlift or tire rotation you master, any good model has to connect the seat of your intellect to the seat of your pants.

We’ve simplified my sprawling spreadsheet of numbers and come up with a more workable formula. We’ve added some interactive graphs that respond not only to price, but to fluctuations in the demand curve (yes, it curves now), and to what degree our dice are correlated. The model now tells us things we never asked it to, and that’s a sign the professor says shows the model’s true value.

Along this journey, I’ve also developed a relationship with Matt Lindsay, president of Mather Economics, a firm that has helped publishers of all kinds think about how they price subscriptions. On the question of how to address ad blockers, what’ll be most valuable to publishers is Mather’s Listener tool, which will value an audience in precisely the way that’s required by our model.

You have to know what your audience is worth, and then think about how you’ll treat them.

This is the question publishers have to ask themselves before offering them ad-free services. The answer will likely differ for each publisher and each audience.

Ad-free pricing depends on your audience’s ad value distribution and demand.

These days, I’m excited every time we find new ways to illustrate otherwise complicated relationships — as if I was a kid tacking a new fin on a balsa-wood airplane. My fellow fellows might think I’m a bit crazy every time I show them the latest version and ask them what they think.

But we’ve started to talk to publishers.

Ideally, we want to share data and resources with those who are interested in this way of thinking and what it could say about their situations. We’d also like to learn from what they’ve tried vis-a-vis ad blockers and see where we can tack on new flippers or gears.

Ad blockers have opened a new frontier in the battle for survival of journalism in the modern age and that may not be a bad thing. They have put pressure on the ad industry to reform, and initiatives like Google’s Accelerated Mobile Pages and Facebook’s Instant Articles are attempts to right the wrongs that led so many to turn ad blockers on in the first place.

But until those systems really take off and can transform the Internet into a friendlier place to go without so many intrusions on our privacy and attention, I think we as journalists have a duty to serve our audience’s demand for ad-free.

You might not know what will happen on every roll of the dice. But just like in Monte Carlo, if you price each bet correctly, in the long run, it’s the house that wins.


Ryan Nakashima is a JSK Fellow at Stanford University studying the feasibility of ad-free subscriptions and their design. Before joining the fellowship in September, he worked as a media and technology reporter for The Associated Press in Los Angeles. He can be reached by email at rnakashi@stanford.edu and on Twitter at @rnakashi .