Thoughts on media business models

Facebooktwittergoogle_plusredditpinterestlinkedinmail

As the smoke begins to clear from the shut-down of my former employer and the rubble that was left behind starts to take shape, a number of nagging questions remain. That’s an understatement, of course; pretty much all that remains are questions, some financial and some emotional. But from my personal perspective, one of the most nagging is: Does Gigaom’s failure mean the model the company was based on — a three-part structure composed of advertising-supported editorial, events and subscription research — is fundamentally flawed, or was it just poorly executed?

A number of news articles, including one by Peter Kafka at Re/code and another posted on Tuesday by Farhad Manjoo at the New York Times, have described the debt and related cash-flow requirements the company faced, which appear to have become too onerous for the board and the company’s creditors to stomach any further (more details have emerged since).

Although revenues at Gigaom’s research arm were growing, what seems to have happened is the gap between those top-line revenues and the bottom-line cash flow or profits of the division grew too wide, and investors stopped believing that the company would ever get there, or weren’t willing to spend any more to fill that chasm. And without a plausible road to profitability, there was no hope of an exit that would justify the $40 million in cash and debt the company had raised.

1-XJU5HSWl0opWukEiWQMZjA

So was the failure of the company a result of a fundamental flaw in the three-legged research/events/editorial model? I don’t think so, despite the rather tangible evidence to the contrary — evidence I am forced to confront every time I look at my dwindling bank balance. I didn’t just champion the Gigaom model because I happened to be working there, I actually believed (and continue to believe) that it can work, and that if done properly it can create a much better foundation for a media company than just relying on advertising.

I mentioned the Economist in my earlier post, and it continues to be the ne plus ultra of this approach: it has an editorial product that is supported in part by advertising and in part by subscriptions, but it also has a highly-regarded research arm — the Economist Intelligence unit — that contributes substantially to the bottom line, and it does events and offers its members other benefits as well. Politico is pursuing a similar model, offering a free editorial product combined with a subscription-based premium product (Politico Pro) that includes events and other benefits for members.

The principle behind this model is that you bring readers in through the front door — the free, ad-supported editorial side, which hopefully carries its own weight (as Gigaom’s did) but doesn’t make a lot of extra money — and you hope to create a relationship with them based around your expertise. Then you (theoretically at least) monetize that relationship through other means, including events and subscription research. In some sense, the editorial product becomes a loss-leader for the rest of your offerings, a way of finding new readers or customers.

paid_content_2745

So what failed in Gigaom’s case? Former VP of research Michael Wolf argued in a recent post that the research arm lost its way at some point, and decided to focus on high-volume (but relatively commoditized) products like webinars and sponsored white papers rather than developing a deep expertise in subject areas — and it also targeted large corporate clients, many of whom apparently didn’t renew after their initial trials, instead of focusing on individual buyers and growing more slowly.

This might be naive, but I still believe that the Gigaom research unit and the model it was part of could have worked, and in fact arguably did work for a number of years. From what I have been able to piece together, it sounds like spending — which was undertaken in an attempt to grow that side of the business to a scale that would make it worth all the millions that VCs had invested in it — just out-paced the actual cash coming in. That’s arguably a flaw in execution, not the model itself.

The bottom line is that I don’t think the death of Gigaom should be taken as evidence that the model itself doesn’t work. Is it difficult to execute? Sure. A guarantee of success? Hardly. But I think there is room for another media entity to make a go of it if they wish to, and hopefully we will see more giving it a shot. Advertising and traditional news paywalls are not the only options.

Comments (4)

  1. Rick Waghorn wrote::

    Matthew,

    You kindly favd a piece I wrote the other day; and I still think it is worth a revisit.

    There is something fundamentally wrong with an ad model – and this is not of Gigaom’s making – that delivers less than $200,000 rev per month on 6.5m monthly uniques passionate about one niche.

    How many events or research teams did the newspaper industry run in its pomp? They had no need.

    There appears to be this industry-wide resignation to the fact that the ad model is f*cked; and yet there goes the Guardian leading the Pangaea Alliance into the panacea of all evils that is, apparently, programmatic and RTB.

    What if Gigaom’s lasting legacy was to be a canary in the coalmine? The first warning shot that the whole ad tech marketplace in its current bot-led, top-down direction is increasingly less and less fit for purpose?

    Best

    R

    Saturday, March 21, 2015 at 2:26 pm #
  2. Mathew wrote::

    Thanks for the comment, Rick. I agree that the advertising market is in free-fall in many ways, but I don’t think this played all that large a role in Gigaom’s decline — our editorial side was actually pretty close to break-even, I think, in part because we got pretty good CPMs. I think overly optimistic growth projections for research were the biggest part of the failure. But I agree that with an audience of 6.5M uniques, it should be possible to make a living from advertising — that doesn’t appear to be the case any more.

    Saturday, March 21, 2015 at 2:34 pm #
  3. Rick Waghorn wrote::

    The other point that has crossed my mind of late relates to the spin-off effect of sites like GigaOm having to seek research and events dollars to make up for advertising’s growing shortfall.

    And I hasten to add that it is my understanding that Gigaom prided itself on not following this particular herd…

    … But part of the event dollars comes from charging speakers to speak; sponsored slots at $2,590 or whatever apiece.

    The net effect was to ensure that it is all too often the legacy incumbents who are hogging the stage as it all comes off the corporate Xs.

    As Adam Tinworth put it, at all too many media conference events the future isn’t invited.

    All too often because the future can’t afford it.

    And so the industry stagnation in terms of fresh voices and new innovation merely deepens. But the event coffers overfloweth…

    R

    Saturday, March 21, 2015 at 3:57 pm #
  4. Stuart wrote::

    I get The Economist reference – which you’ve mentioned before – but don’t forget that the publication itself is really expensive to purchase. I mean, I bought one at the news stand before a flight the other day (I – like many Gen Xers it seems, per a bunch of media research I’ve looked at, love The Economist) and paid something in the area of $9. Even subscriptions, online or off, aren’t cheap. So, yes, without a doubt their Intelligence Unit and Events do well as a part of the mix. But they have still managed to effectively ring-fence a bunch of specialized, well-presented content (in the form of the publication or app) and obtain premium pricing for it. Which, in this day and age is both not done easily and calls the “content for free, research and events make the money” approach into question.

    Wednesday, March 25, 2015 at 6:09 pm #