Thinking about a paywall? Read this first

If you're thinking about charging for content, this high-quality infographic could save you from making a big mistake:

What is this? It's the general shape of your typical newspaper website's user behavior over the course of a month.

This is a not a graph that your typical Web metrics system offers up without a struggle, but if you can force it to reveal this data, it's eye-opening. I first pulled this sort of curve out of a Tacoda system half a dozen years ago, and it changed a lot of my thinking.

On the left side, you have awesome reach, both in-market and (as previously discussed) out-of-market. As you get better with SEO techniques, that spike gets even higher.

Most of those visitors come once or twice, probably following a link from a search engine or another website. They're looking for something very specific. They find it (or not) and leave.

Then the number drops like a rock. Hardly anybody comes five times in a month.

But over on the right side you have an interesting little lump.

It's small, but that's deceptive. That lump accounts for a big chunk of your website's traffic, because the people in that lump visit several times a week, maybe every day, maybe two or three times in a single day. When the visit, they consume maybe eight, maybe a dozen pages.

These are your loyalists. Compared with the overall population of visitors, they're far more likely to live in your market. They're keenly interested in your content, highly engaged in local life, and solid gold prospects for your advertisers.

Now here's what's wrong with a paywall. If you're trying to persuade people to give you money for your content, it's the wrong tool. It's like cutting butter with an ax handle. You're just going to make a big mess.

These guys over on the left side of the graph are not going to pay. The best you can hope for is a percentage of the people in the lump on the right. And in terms of a nose count, that's not a big number. So rein in your dreams of paid-content riches.

But wait, it gets worse. Those people over on the left side? Some of them are out-of-market lookie-lous, but some of them aren't. Your future is somewhere in that spike, and your paywall just told those people to eff off and go away. Your goal should be to persuade them to become regular users, not beat your chest and make demands.

Now, here's something that people from the online side of the fence can learn: A lot of people who have been thinking about paid content are discovering this curve. And those who do are coming to favor a completely different paid-content model, one that would leave the left side of the graph unmolested, and concentrate on persuading the people on the right side of the graph to pitch in some cash.

Several years ago when newspaper websites were all setting up registration systems, most of them did it completely the wrong way: you had to register to read anything. The one system that did it the right way was McClatchy's, which had a (configurable) threshold. You could read up to N pages per month. After hitting the limit, you'd be challenged and asked to create an account.

Random visitors were unmolested. Loyal visitors were asked to provide demographic data. Most did so, and those of us who actually dug through that data know that it was remarkably honest and accurate.

Journalism Online, the startup that's trying to persuade newspaper publishers to charge for content, is currently pushing a paid-content version of that model.

What should be the threshold? How price-sensitive is that group on the right side? Can you get significant reader revenue from that group without damaging your advertising business? Nobody knows, really, and surveys aren't likely to tell you.

It may not work at all. But it has going for it this one really important characteristic: It's looking for opportunity. In a discussion that has been dominated by ridiculous, absolutist declarations, that's a welcome change.

It will take some experimentation. Given the risk, I'm eager to see somebody else do it. Change is good; you first.


Steve, This is consistent with the marketing concept some call reversing the funnel. Instead of starting at the widest part of the funnel (the left side of your graph) and generating as big a universe as possible, "hoping" some will adopt the brand as their own, "reversing the funnel" starts at the smallest part of the funnel, winning the credibility of the most receptive in the category (usually already active participants). It's kind of funny that this bottom up approach is called "reversing the funnel" since it really is the old-fashioned selling. The top down approach is a relatively recent phenomenon, dating to when the 3 networks dominated television viewership. Then, it was possible to reach 80% of Households in a market in real time, by "roadblocking" a timeslot with a commercial spot on all three network affiliates. If you got the message right, you could predict that about 15-30% of them would be curious enough to go to the store and try it. If the product delivered what was promised, from 60 to 80% would buy again. By the late 1980's when cable penetration had reached critical mass, it became impossible to reach 80% of anything. The more spots you buy, the more frequency you get. The advantage of bottom up thinking is that when the product is designed to engage the loyal, paying customer you know instantly if it works or doesn't and can even ask why. The more satisfied and invested your core customers are, the more likely they are to talk about you and influence more of their peers (on the left side of your graph) to subscribe, too. That's a much more efficient way to expand reach and results in higher conversion rates. At Comradity, we've learned a lot about how to increase engagement and accelerate the influence of loyal customers on their peers. If your interested in learning from us, go to our homepage and look at the prezi: to see what we've learned. Katherine Warman Kern @comradity

There is a difference between paywall and paid content. If you have a lot of free content online AND some content you need to pay for it is NOT a paywall You should read this as well There is no way commercial media can survice only on advertising! @pernillet

Please define or clarify what you mean by "typical newspaper website"... are we talking the local paper, like the Galena Gazette, or national like USA Today etc? I think the difference is important, though the result *might* be the same.

Steve, where did this graph come from? Is it based on real data or extrapolation / speculation? We've had a paywall, at a small community paper since 2005. The link below shows a side by side comparison between our PAID site and a neighboring FREE site of similar size. You will notice that the frequency (visits per user and pageviews per user) is actually higher on the PAID site. Paywall's may not work in highly competitive markets but they sure as heck work in ours.

Joe: The chart is obviously a work of art (ahem) but it reflects real-world data from our newspapers. As for the frequency issue -- your frequency numbers should be high on a paid site. You're locking out the casual users. A simple paywall kills reach and growth, but among those who want your content badly enough to pay, the usage frequency should be high.

Pernell: Commercial broadcasting has been a money factory since the 1920s. We shouldn't approach this issue with predetermined articles of faith, but rather open questions about what works.

Michael: Local newspapers. We have over 1,400 local dailies in the United States, versus three national newspapers. I actually don't regard WSJ/NYT/USA Today as being in the same business as the rest of us.

Though I'm still not convinced that a paid-content model that uses this graph to inform its logic will be successful (particularly if the content is just repurposed general newspaper fodder), I love the logic expressed here. Only thing I'll add: Rather than challenging loyal readers to provide information to keep reading, use returning visits to recognize your loyalists and then make them a value proposition: In return for X, we'll give you Y. I suspect that approach will produce better results.

Thinking a little more broadly about content - not only reporting, but other things people want to consume and share - it makes sense that this model that worked so well for sites like Flickr would work for news sites also. It just makes sense. Whereas a paywall divides the audience into those who will buy and those who won't, this way targets the people who likely will, but also leaves the rest in the maybe category, instead of making them definitively nots.

Thanks for the follow-up. To keep it in context, you said: (This is in relation to) "Local newspapers. We have over 1,400 local dailies in the United States, versus three national newspapers. I actually don't regard WSJ/NYT/USA Today as being in the same business as the rest of us." In that case I'd be mindful of people making arguments for the larger papers based on this - and I think they are already.

I don't have any data on the national sites. Their usage curve might look the same, or it might not. If it does follow the same pattern, it might be helpful in analyzing their opportunities -- but there are many differences from local media. One is radically higher competition. Another is that the loyal and incidental users of a site like USA Today would seem to be fairly interchangeable in terms of economic value per page viewed, which is not true in the local space.

I don't mean to be argumentative, but I still don't see the benefit for the small paper to remain free. The following link shows LIVE stats comparing our PAID site to the neighboring FREE site. Yes, the free site has more pageviews, but the curve looks nothing like the graphic in your post. And we are able to charge DOUBLE the CPM rates of the free site. Our paid site is by far a more respected and valuable to local advertisers.

Good post and great graphic. It remains a constant no-brainer really. Auntie Beeb will almost certainly remain the free source of propaganda it always has been and the only people likely to subscribe to news sites will be those needing information from specialist publications - lawyers already do this with Lawtel and other professions have their own versions. Free sources of news will always be found elsewhere and whilst bloggers are generally discredited by mainstream media, the quality of some of them has made increasingly more people question the "objectivity" of proscribed stories in our news sources. And mainstream media's love affair with celebrity and lifestyle is surely not something many will consider important in these economically austere times. A paywall "on-line newspaper" will become little more than a web site just like any other. Nobody says "Oh, I take the Times" these days, or is that loyal to the predictably dreary Guardian, shoestring Independent, etc. Mainstream media is falling apart. I feel sorry for all journalists having a hard time and who are trying to adapt (and I am one), but no one else wants to go back to the information distribution low technology of the 19th, 20th and early 21st Centuries. To try would be as futile as smashing a few printing presses. Ordinary people are voting with their keyboards or mobile phones when it comes to national issues and finding that the traditional models - however they are dressed up digitally - are dull, predictable, iconoclastic dinosaurs. let's also recognise that much traditional investigative text/photo journalism disappeared long before Google arrived. And local newspapers are missing both the point with their own paywall aspirations (e.g., ala Johnston Press) and the opportunity for local print sheets. Local people still like local news in print form, and advertisers can benefit from this. A TMP 2007 poll found 77% of readers trust local paper ads more than those they see on TV or the internet. Events listings, property ads and all other local material can usually be found on numerous blogs and independent sites ... and so who will readily subscribe to get over a paywall for local "news"? I actually hope the BBC and entire mainstream media do erect paywalls and ultimately flounder. Then, at last, false mainstream paradigms will wither and new opportunities can flourish. Viva la Revolution! (Even if it is a painful evolution for some of us in the media sector.)

I love reading your blog and your thoughtful commentary on digital media. I do, however, disagree with this post and those by others that I've read proclaiming that newspapers can forget it -- there is no hope for selling their product in today's marketplace. And that's because I think we're having the wrong conversation. Can newspaper companies take the product they are producing right now and successfully sell it online -- to readers or advertisers? No, in most cases, I don't think so and that's proven to be the case in many oft-cited examples. Newspaper managers, for decades, have had the luxury of not worrying about whether they were delivering editorial quality, results for advertisers or customer satisfaction on either side of the company. Their customers did, however, need a little bit of something they offered and since they were the only company in town offering anything like it, they sucked it up and shelled out the cash for it. Now people have been given more choice, and the fact that they are unwilling to pay for most newspaper products doesn't mean they won't ever. It doesn't mean they are bad/stupid people who are destroying democracy. It does mean they know a product not worth their money when they see one. I believe that if people aren't signing up to pay to access your newspaper online, you need to look at it like any other business would look at a product that wasn't generating the sales they thought it would. Find ways to improve it -- I mean REALLY improve it. Examine your market. Talk to your customers about what they value (specifically) in the many services you provide. Start having conversations with your staff -- from the top to the bottom -- about what is working, what isn't and how you can fix what isn't. And commit to making big changes, even if it means building a new system and/or new expectations from scratch. We can't keep trying to sell our product the way we always have and expect the world to change back to our liking.

This is great chart - can be used to decide whether to pursue paid/registered content, or how to do so if that decision has been made. I have seen the same curve for local newspaper site/s and national niche sites (both consumer and b-to-b), although I seem to remember a larger "loyalty hump." (in one case, it might have been 20% of users come 1 to 4 times per weekday.) I also liked Tacoda for this measurement because it tried extra hard to "unduplicate" unique users, which made for a smaller audience but a larger loyalty hump. This data gets really interesting if you can filter your stats for local users only, for print subscribers vs. web-only customers, or even for usage of a paid digital edition if you've got one. Heck, I suspect retails stores would have a similar chart for window-shoppers, once-a-year-back-to-school-bargain-hunters and the people-with-a-store-brand-credit card. The shape of the humps may change, but the pattern is similar.

Of course not! Come on, they can't be, the x-axises are totally different! Your x-axis is time, in days. The x-axis in Steve's graphic is "frequency", meaning how many times a user visits the site in one month. So, you're comparing apples and oranges! Didn't you notice that? Uh, this makes me guess you're not the administrator of those sites, right?

Steve, Good post. For an alternative take on the same ideas see my post at BTW this is not speculation - it is what the Financial Times has been doing successfully for a while. Best,

First, I don't exactly know how you generated your graph, but it sure looks like visitor loyalty. In my case (Google Analytics) the visitor loyalty look roughly the same as your graph, but... the last 6 rows show 9-14 times, 15-25 times etc. These accumulated results are the bump in the graph. It's a bit of a distorted view because actually there is no bump in the graph. Second, the GA visitor loyalty shows visits, not visitors. So on the left side of the graph are visits from new visitors and on the right side of the graph are visits from frequent visitors, but it doesn't represent the number of frequent visitors! Also see: This doesn't really affect your theory about the paywall, but it is very important for people now looking at their visitor loyalty metrics and possibly misinterpret them leading to false conclusions. You wrote: This is a not a graph that your typical Web metrics system offers up without a struggle. Maybe you could give us some more insight in how to generate this graph with true visitORS and frequency. Grtz. Jeroen