virtualeconomics

Facebook gets ready to mop up the rest of the display ad market

Facebook just announced a new messaging system that isn't quite email or IM but something inbetween. There's plenty of coverage of the press conference out there so here's my first thought for now - this is about the display advertising market.

Facebook already serves 25% of US online display ads. Commercially, that puts it in direct competition with the two gorillas in that space - MSN and Yahoo. Both of the incumbents have a lot of advantages selling online display inventory - a mix of premium content with huge quantities of cheap network inventory, their own third-party ad network, behavioural targeting data and tech plus the best buyer relationships in the business. But those fortunes were founded on - and still rest on - the mass inventory that comes from running the world's two biggest webmail services.

The question to ask about Fmail is not about the tech or the user experience but about the money, the money, the money. FB, that is already taking over the display ad space, just announced a product that could cripple the foundations of its two main competitors. That's what Fmail is really about.

(This post was written on a phone and has subsequently had the links added.)

News Corp's paywall is about News Corp, not the Times

Round about the same time as it makes a bid to acquire the whole of BskyB, News Corp has also erected a paywall around the Times. Analysis of this move to date has focused on what this means for the Times qua the Times, and what the lessons are for newspapers generally, and produced some valuable insight as to the meaning behind the fluffy numbers. But I believe this move has almost nothing to do with the Times and everything to do with News Corp's wider bundling strategy, and that considered in this light erecting a paywall around the Times was a sound strategic move which has the potential to return the Times (not just the website) to break even.  

Considering that the Times lost £87.7m last year and BskyB recorded profits of £855m on revenues a shade under £6bn, the obvious question for anyone holding both businesses is not "how can we make the Times make more money?" but "how can we make the Times make Sky more money?".

Sure, while a BskyB subscriber is worth £508 per year and someone buying both the Times and the Sunday Times every day is worth £312, the reality is that not every Times reader subscribes or buys every day (fewer than 150k subscribe, from a circulation around 500k at the Times and 1m at the Sunday edition); the circulation figures for the papers fell 15% and 10% respectively year on year; and in any case that's all just talking about the print audience. The online audience, which pre-paywall stood between 3m and 6m depending on which web analytics provider you believe, is estimated to have been worth £10m-£20m in ad revenues last year. That works out at an average value of a visitor to the pre-paywall Times website 0f about £0.30 per month or £3.33 per year, which in my experience is certainly the right order of magnitude and likely to be in the ballpark. In context, Timesonline visitors were about 1/100th as valuable as a print edition loyalist, or 1/150th as valuable as the average BskyB subscriber. Losing 90% of them, and the attendant £10m-£20m in online ad revenues, is a shame, but on an ARPU calculation for News Corp it simply doesn't matter. 

So the paywall isn't about monetising the Times (the best estimates hold that the paywall is generating revenues between £4.8m and £5.5m pa,which in the context of £6bn pa isn't worth the trouble of counting it). BskyB has about 10m subscribers. Of those 2.6m take broadband, 2.4 take telephony and about 2m (21%) take the lot. The business added 1.4m subscribers last year, but with a churn rate of 10.3% that only added a net 400k to the total - or, to put it another way, last year 1m customers worth £508 each or half a billion pounds walked out the door. The Times paywall, therefore, is not about directly monetising the Times but about bundling the Times with News Corp's other services, notably broadband but also TV and telephony, to decrease churn and add customers to the valuable growth part of the business.

The three major UK providers of bundled services are already operating a confusopoly - a search on the major comparison site for a TV/broadband/phone package at a sample London address indicates most clearly the presumably deliberate impossibility of comparing like with like. For £190 a year is Virgin's "medium" TV package with 40 channels and 10mb broadband a better deal than Sky's "basic" TV package and 20mg broadband for £258 a year? Clearly and deliberately this is a decision that cannot be made on price alone, and adding the newly-paywalled Times to that bundle makes it possible for News Corp to retain or add more valuable subscribers to its core business. 

Looking at the numbers, the pre-paywall Times generated 41m pages from 6.4m users; that's now down to 4m pages from 2.4m users. Elsewhere this has been interpreted as a loss of 60% of the audience - I think the lost audience figure is far higher, since pages are down 90% and the way traffic is counted almost certainly includes as "visitors" people who bounce off the paywall. The real audience, actually reading pages rather than bouncing off the paywall, is probably reflected by the page views figure and is down from 6.4m to around 640k. That's 5.8m who used to read the Timesonline and now can't.

At this point it comes down to marginal decisions. We've seen that you can't compare the TV/bb/phone packages on price - the confusopoly makes it impossible. So at the margin, how many people need to decide that on balance a free sub to the Times online is worth switching from Virgin or BT to Sky for this to work? Top-end estimate for lost online ad revenue is £20m; knock £5m off that for online subs under the new model and the paywall is losing £15m a year. With BskyB running an ARPU of £508, if fewer than 30,000 people make the decision to move to (or stay with) Sky because they get the Times online bundled in, News Corp's paywall initiative breaks even. If 200,000 do that's the annual losses of the newspaper operation covered too. In context, the UK pay TV market is about 14m people (own calculation from Sky results plus here and here); the broadband market about 19m; and Sky already churns a million people a year. 0.2% of the UK's broadband market needs to switch to, or stay with, Sky this year for this to add up. Assuming I'm right and they are going to bundle, the maths at the News Corp level work out just fine for a paywall around the Times. And I hear rumours that the Sun and the News of the World have websites too.

 

Addendum. This is obviously back of an envelope stuff. Without access to Sky's internal figures it can never be anything else. There are complexities underlying these basic numbers - for example there would be some, perhaps considerable, deadweight loss in both directions if the bundle was as simple as I've suggested. No accountant would really attribute every penny of a marginal decision to spend £508 a year on a package of TV channels and connectivity to the presence within the package of a free online newspaper. But I believe it's directionally right, and the picture looks even rosier when you consider that if News Corp really does pursue this strategy the audience for the Times online is going to increase dramatically; it will continue to be made up entirely of paying customers for whom News Corp has credit card details and all sorts of deep behavioural data; and at that point the people who are currently eschewing the site as an advertising medium might be willing to reconsider. 

Detargeting

Here's a fascinating idea - adding banner ads to corporate emails. Let's briefly think through what's wrong with this. Take a highly targeted, personal, one-to-one comms medium (email) and dump a mass-marketing branding solution on top of it (banners). You've got all the elements of an unstrategy right there: a wasted opportunity (if you want to send commercial messages out with your corporate emails it's possible the sender knows a little bit about the recipient, at least enough to tailor the message), a nuisance cost that diminishes the utility of a previously adequate channel and a fairly serious chance to harm the credibility of both the advertised brand and the email sender. Best result I can see - corporate email gets used a bit less because it's slightly less useful and people spend a bit more time face to face. Or we get an AdBlock add-on to email that strips the banners out. Anyone working on that yet? You could probably sell it as an enterprise solution...

ebuzzing: PayPerPost with (much) less evil

Once upon a time, so legend* has it, there was a French foreign minister who used to make all of his decisions in the genuine and sincere interest of France and then only afterwards try to work out whether there was anyone who would want to bribe him to do what he had decided to do anyway. I've always rather admired the idea. Taking money from the criminally corrupt for doing what you were planning to do anyway is practically the perfect crime.

I mention the story because today I first heard about Ebuzzing, a new blog monetisation tool that aims to pay bloggers for posting about Ebuzzing clients.

We've been here before, right? Obviously this is just the old evil of PayPerPost in a slightly different hat.

Actually, I'm not sure it's that cut and dried. Talking to the company today, there seem to be some important differences. Colour me hopelessly credulous, but this looks like a genuine attempt at PayPerPost without the evil.

First, disclosure is required - you've got to admit you're getting paid for the post. Second, there's a search "no follow" policy. (This is clearly just to stop Google clobbering ebuzzing like it clobbered PayPerPost but still - it really does mean the posts don't create a lot of spurious search mojo.) Third and most importantly, according to Dan Levy at the company:

"We insist that our bloggers post only about things that they genuinely are interested in. If it's not a product or service you wish to draw to your readers attention, you don't back the advertisers campaign. That way even when you are paid for blogging about something, it's because it's something that you actually want your readers to know about."

See, that's the bit that I think might work. To return to my French foreign minister analogy...once he's decided the course that's really in the best interests of France, why shouldn't he find someone to bribe him for it?

Of course, there's lots of possible answers to that. Motivation is complicated and once you know there's money on the table you'll always be second-guessing yourself as to whether it affected your judgement. And if it's problematic to prove to yourself what in order you made the decision, it would be horrendous to have to convince a third party. (Imagine for a moment committing a bank robbery for the sheer thrill of it and with the sincere intention of returning the money the next day, but being caught by the police before returning it. Somehow it seems unlikely you'd be believed.)

And yet...ebuzzing seems to me to be making a genuine attempt to solve all of those problems. Dan at ebuzzing again:

"We read all articles and will to the best of our ability decide whether a blogger is feigning interest if it appears that way. In the few cases where we have felt it was happening, it was fairly obvious, and we asked the blogger in question if they really were interested in the product. Most of the people in these cases decided to drop the piece and it therefore never left our platform (and hence never went on their site)...If bloggers feign an interest in something then it is obviously likely to be very apparent to their readers, likely more so than it will be to us. It is of course possible that some bloggers who don't blog regularly will decide to cash in if they can. We cannot reach into their heads and see what they really think. But we've found so far that it's very hard to fake that kind of approval in a blog post without sounding ridiculous, and even harder to fool your regulars. The proof is always in the editorial pudding, for everyone to see."

It feels...plausible. When he first slammed PayPerPost Jason Calcanis said "the currency of blogging is authenticity and trust... you pay folks to blog about a product and you compromise that." And there I think lies the solution. I trust the bloggers I read to be authentic. If they occasionally tell me they've decided to write about Apple or Coke or some cool new start-up and then found that ebuzzing is willing to pass them a few quid too, I think I'd be prepared to believe them. I doubt I'd trust a stranger's ebuzzing-flagged post. I doubt I'd trust even a blogger I knew quite well if he started doing it more than very occasionally. But on the handful of occasions when I've sat down to write about the excellent customer service I got from my bank or how much I like an online service I use - perhaps three or four times in the two years I've been writing this blog - I think on balance I'd have been happy to take money from the company I was writing about, given the protections ebuzzing appears to provide.      

*Or if not legend, history graduates who don't have the original source to hand

Unintended consequences of ad targeting

As ad targeting gets better and better, I find more and more intended uses for the ads.

Last year I commented that some friends and I had found the easiest way to be alerted when ski chalets became available at the resort we wanted to visit was to knock up a wiki page with the name of the resort and a few words about skiing, run AdSense over the page and then wait for the relevant ads to be brought to us. As ad targeting gets more precise (and accurate) it becomes trivial to simply design a page so that the right ads come to you. AdWords are not just commercial messages: manipulated the right way, they're a tailored newsfeed about things you want to buy. 

Yesterday I found what I'm confident will become another use for ad targeting. Pondering attending a pub quiz, I had a look at some web pages that talked about the pub in question. The ads alongside those pages were all for distressed-credit services such as LogBookLoans, debt consolidation and other financial services clearly designed for what marketers would euphemistically call the C2-D/E audience. This gave me pause. If the ads were targeted correctly (and increasingly they are), perhaps this wasn't the event or the venue for me.

We are regularly warned that our Facebook (etc) pages reveal too much private information about us. But that's just the information we voluntarily put on there. I increasingly wonder, as ad targeting gets better and better, what sort of information people will be able to glean about us and our social circles from the sort of ads that targeting systems choose to appear alongside our profiles.

Update: in comments Ross Parker points to some entertaining examples of ad targeting gone wrong at imedia connection   

Economics of the click

danah boyd looks at who clicks on online ads. Since click-through rates are miniscule (and falling), and since geeks block ads anyway, it’s an interesting and important question – who is actually clicking the things?

According to one study by AOL, “we learned that most people do not click on ads, and those that do are by no means representative of Web users at large...Ninety-nine percent of Web users do not click on ads on a monthly basis. Of the 1% that do, most only click once a month. Less than two tenths of one percent click more often. That tiny percentage makes up the vast majority of banner ad clicks.” danah goes on to speculate that this tiny group of clickers are unrepresentative of the online population as a whole - that they are essentially the same audience as responds to direct mail and telemarketing, i.e. a lower income, lower education, non-metro, female-weighted subset of the total online audience. 

Which has some interesting implications for online media economics. Publishers sell their ads on the back of demographic data. Much is made of the fact that site x has an audience composed of 75% ABC1 women and site y an audience of wealthy urban men. But if those people don't click the ads - if in fact it's just the same subset from the lower end of the socio-economic scale that's doing all the clicking, irrespective of the site, that's a potentially huge problem for display advertising.

Or perhaps there's no problem. The value of an ad is not just the click and much advertising is bought to build brand rather than generate a direct response. danah's conclusions are largely speculative and the post is upfront about the need for more data. However, it's an interesting question and one that merits further investigation - if click-through rates are so low, just exactly who is it that creates those few clicks?

AdReady misses the opportunity to create an ad marketplace

News from TechCrunch today that a new ad solutions start-up, AdReady, has missed a trick by offering everyone the tools to create professional-looking display ads. What trick? Well...currently people can go to AdReady to create a display campaign, and use its data on the historical performance of various creative formats to choose the best ad for the job.

Nick Gonzalez comments at TechCrunch that "I imagine an extreme reliance on the data will produce a lot of generic advertising", but the missed opportunity is larger than this. It's a shame that people will be using this merely to create generic banner advertising, but the real loss is that there's no opportunity here for people who want eye-catching display ads to connect with people who could design them. (A glance at Amazon's Mechanical turk project shows us the incredible cost-effectiveness that it's possible to tap by opening up "human intelligence tasks" to a competitive marketplace.) As Robert Young once said of social networking tools,

"Think of this way… what if “American Idol” had been produced solely by the capabilities of the contestants themselves, without the expertise and talent of the show’s producers, directors, writers, etc. As talented and entertaining as the contestants are, the resulting production quality, the level of emotional engagement, viewership/ratings and monetization potential of the full package would likely be far inferior to what we all see on the air today."

Giving people the tools to create their own media can result in extraordinary results, but it has its limitations, and there seems little reason that AdReady (or a solution like it) can't tap a market for design expertise rather than merely hand over the tools for amateur ad creation.

Update: AdReady's marketing guy points out (within a very few hours of the original post going up - these people are on the ball) that there's already an element of this on the site and that "we expect to make the ability to design creative for advertisers through the AdReady system more available over time, as I absolutely agree that the collective ideas and talent of designers around the world is not something that any single company could hope to match." So perhaps AdReady will become a creatives marketplace, which is where the big prize will be.

Year Zero

If you haven't seen Year Zero yet, you really should. Depending on who you ask it's a game, a virtual reality or a marketing gimmick for a Nine Inch Nails album of the same name that was released earlier this week. It was originally found by fans on a website they identified by decoding highlighted words on a t-shirt (Wikipedia), and further plot developments were left on USB sticks in the toilets of NiN concert venues. The album (and the game) are about a dystopia fifteen years in the future and include dedicated websites for such in-game characters/plot devices as the 105h Airborne Crusaders and the First Evangelical Church of Plano.   

As well as being, obviously, gloriously strange the whole thing goes way beyond the current state of the art in marketing or advergaming or any sort of cross-media promotion - NiN frontman Trent Reznor comments: "It's not some kind of gimmick to get you to buy a record - it IS the art form... and we're just getting started."

The only thing this really reminds me of is "the footage" in William Gibson's Pattern Recognition (not read it? Do. There's a link right there you can buy it from). Yet another "we are living in the future" moment.

Innocent: pretending to be evil

Some companies make a lot of noise about not being evil. Innocent, the guys who make the (nice, expensive) fruit drinks, go one better than not being evil. They pretend to be evil. They riff on the whole being an evil corporation thing, just to make the point that they're not one. What do I mean? Well, look at their ad at the bottom (click it, or here, for a better image).

See the asterisk? Usually, an asterisk in an ad means "we're lying to you. Follow the asterisk to see the many, many qualifying statements that we're legally obliged to add to our wild claims". When Innocent does it, the small print is just a joke - "it's a tight fit. Like spandex trousers". That's great branding. That's the confidence to pretend to be evil, and trust your customers to get the joke.

Beyond the words

Yesterday saw an almost hysterical call by The Philadelphia Daily News for Craig Newmark, whose free classifieds model is demolishing the most valuable classifieds sections of newspapers, to set up "Craig's Foundation" to, amongst other things, "build...a retirement home for any newspaper folks who might somehow see a diminished pension down the road". Wow. Terry Heaton comments with due incredulity that the writer "sounds a bit like a whale oil spokesman casting aspersions on Thomas Edison", Nick  Douglas at HuffPo says that it's like "asking Gutenberg to please fund the monks". Quite so.

Now, I've commented before that journalists are hardly unique is disliking change - dislike of change is a wholly explicable and reasonable human survival trait. It's not really news that some journalists are feeling pressure from the web, or that what was once a successful bundling strategy - the newspaper - has, like the CD and the TV network, found itself unbundled and the parts of the package re-examined for value in light of the digital revolution. Talking about this digital unbundling in the WSJ last month William Bulkely noted that "the internet allows consumers to trim wasteful purchases" and at Slate Jack Shafer summarised his key themes:

"the photographic film industry, encyclopedia publishers, the music industry, and the advertising industry feasted on buyers by forcing them to purchase things they didn't want—prints of all 24 shots from their camera or a whole album to secure one favorite song, for example. "The business models required customers to pay for detritus to get the good stuff," Bulkeley writes. But digital cameras, the Web, iTunes, and search-related advertising have stripped those industries of their power to charge for detritus."

And as I never tire of quoting Jay Small on the fate of newspapers, I'll do so again: "must we live or die by the bundle? I think not. Unbundling and examining each value proposition for its new media potential would be the best way to manage both cost and revenue curves. Jettison those little things a newspaper does because they make sense in the bundle, if they don't make sense apart from it."

Now, what about that unbundling, specifically the unbundling of newspapers, does yesterdays Philadelphia Daily News rant (yes, it was a rant) tell us? Perhaps that one of the wasteful purchases that the internet allows consumers to trim is the journalism component of the newspaper. I've said before that outside the newspaper bundle, hard news is hard to monetise. And I've pointed out (as has the AJR) that the high-minded journalism that goes into newspapers just doesn't seem to be what people are most keen to read once the newspaper package is unbundled online. The writers at PDN seem to find the unbundling of their newspaper's lucrative classifieds section especially threatening, but look at what that threat means - they think, openly, that readers will not pay enough to fund their writing, in isolation, when given a genuinely free choice.

When Knight Ridder broke up, McClatchy bought the old Knight Ridder papers and then sold on a number of them. Two of these, the Philadelphia Inquirer and - indeed - the Philadelphia Daily News, went to private equity - Philadelphia Media Holdings, headed by Brian Tierney who is now both the CEO of the holding company and publisher of both newspapers. The two papers have failed to meet the revenue targets that their new private equity owners put in place when they bought them, causing much talk of substantial lay-offs. Much of that problem stems from the flight of classified inventory, and therefore revenue, to the web, and especially to Craig Newmark's free model. Hence, pretty evidently, the almost hysterical  bitterness of the writers at PDN. (I'd also add, as pure speculation, that it seems likely to me that this latest article presages some big news coming out of Philadelphia Media Holdings soon. But as I say, that's just speculation.)

But let's put the matter in perspective. Craig Newmark has not done a bad thing here, as some very far from disinterested commentators are claiming. By unbundling the classified ads from the news, Craigslist has done nothing more sinister than given the (ex) readers of the Philadelphia Daily News more choice. That they've chosen not to keep reading the paper is an indictment of the journalists writing there, not Craig. So let me put it like this. He gave your readers more choices, and they chose not to read you any more. If people don't want to read what you write, you don't get paid. Them's the breaks.