Red Hat, Linux, consumers, money – do they mix?

Earlier this year Red Hat split its range into two segments - effectively, business and everything else. It did this in the wrong order, as The Register remarked at the time, but subsequently the company has padded out its Enterprise range to consist of Advanced Server, Enterprise Server and Workstation, so here at least it has a complete set.

But what's going on with 'everything else,' the product line Red Hat terms 'consumer'? After speaking to Red Hat marketing VP Mark de Visser last week The Register left with the strong impression that the answer in the short term was not a lot, or more of the same, while the real upheaval will take place in the longer term, and probably doesn't yet have anything you'd call a serious blueprint. Our parting impression, incidentally, should not necessarily be viewed as criticism; Red Hat, in common with the other major distributions, is still trying to figure out how you make money starting from free, and the bottom line is that the enterprise line shows prospects of making money while the consumer one as currently constituted does not.

When talking to us de Visser stressed that he wasn't there to make product announcements, but he was prepared to talk about the company's views and aspirations in sufficient depth for us to make inferences. So note that what follows is largely Register inferences, not a Red Hat roadmap.

Recent history provides a jumping-off point for where Red Hat consumer is going. In (vaguely approximate) tandem with the enterprise/consumer split, Red Hat segmented its Red Hat Network online update service between paid-for and 'demo' (i.e. not paid for). Paying subscribers are intended to get priority on bandwidth during heavy load periods, while the demo ones have to fill in a Red Hat survey at regular intervals in order to keep their demo account active. Funnily enough, talking to de Visser prompted The Reg to try to tidy up its demo account Red Hat 8.0 installation, but attempts several times a day since Tuesday of last week (twice on Sunday - Sunday, for god's sake) have been spurned with a network congestion notice and suggestion we pay up.

This is a radical change from our previous experience, but our inference as to why this might be the case differs a tad from the obvious.

Red Hat has decided to beat everybody with a stick until they cough up for a sub? There's maybe an element of that, but consider the likelihood that the end result of such an exercise would not be profitability for the consumer line. The people who use Red Hat demo accounts largely got the software for free (perfectly legal, goes with the Linux territory, but not obviously profitable for the distributor) and if using Red Hat software for free gets too painful they'll just go somewhere else, they won't pay. Plus, they'll heap ordure on Red Hat as they leave, meaning that the 'beat with a stick' approach gets you no profitability, the mass defection of enthusiasts and a squalid reputation.

So try a slightly different inference. Red Hat currently has one line that looks like it might make some financial sense, and another (which more or less consists of its whole previous Linux business model) that doesn't. Note it's not the case that the free stuff doesn't make money while the paid-for consumer stuff does - really, none of it does. Red Hat makes some sales at retail, but surely not a profit. Distribution costs take a bite out of the revenue, as does the cost of the Red Hat Network sub. Plus, the traditional open source 'twice a year' update must screw up the logistics of packaged software at retail, so while from an accounting point of view you can split consumer into people who've paid a bit and people who haven't, from a strategic planning perspective you're better off just viewing it as one big, unprofitable pile.

In the case of the enterprise range the customer's perceptions and expectations differ, and it's feasible to segment support systems in such a way as to make sure you get paid for them. But the Red Hat Network is just a big pile of expensive pain, and the more users you get, the more it hurts.

Mark de Visser did not tell us Red Hat is therefore putting the squeeze on it, and not having been squeezed at that point we didn't ask, but he was open enough about the expense, and roughed out a few ideas the company has for dealing with it. These seem to centre on an attempt to offload/distribute the bandwidth usage. With its most recent OS update, he says, Red Hat tried to decentralise distribution, so in exchange for access you undertake to do some distribution yourself, and thus a tree structure that spreads the load and cost emerges. This seemed to work, Red Hat seems to like it, wants to do it some more.

It's possible The Register's update problems over the past weeks have been caused by happy bean counters consequently throttling the bandwidth, but we wouldn't jump to that conclusion. We would however presume that an effective throttling over the longer term is an inevitable consequence of Red Hat easing itself out of direct distribution.

Other ideas relate to this. For example, the company is currently planning a pilot, in Italy, of a subscriber magazine that would include updates on CD, and also Red Hat compatible software. Magazines being where The Reg lived for many years, this strikes us as a plausible, quite possibly profitable, distribution channel for free software.

And what was that about Red Hat compatible software? The Reg suggested to de Visser that the inclusion of a question on the subject of third party software in a recent demo account survey was significant. He responded that as the questions were largely what might pop into the head of whoever knocked the survey together, one should perhaps not read too much into them. But we got talking about third party distribution and marketplace ideas, and it does seem clear that this one has a leg or two.

We infer again, as so often. A subscription software marketplace such as Lindows.com's Click N Run is, we reckon, an idea that could work, but Lindows almost certainly doesn't have the weight and reputation to pull it off. A major distributor however would stand a far better chance of attracting third party software in sufficient volume to attract users who, for a little bit of money could get easy access to apps they knew would install easily and run on their distribution. It could be run out of the subscriptions, it could be subsidised by app vendors giving the host a percentage of the revenue, or it could be a mixture.

It's not, you'll note, exactly a getting out of distribution strategy, but it's a route for supporting the cost of distribution, and open source needs, as it matures, marketplaces.

Tellingly, de Visser doesn't see retail - as in boxes on shelves - as being somewhere Red Hat is going to be in the future, so building paid-for electronic distribution channels becomes vital for the company.

But to distribute what? tricky one, that. While there's a clear market for Linux servers in business, and at least the prospect of a market for 'business workstation' products, the state of consumer Linux is somewhat murkier and - oops - haven't you just tagged the whole of your previous operations 'consumer'? Actually no, we didn't ask him that one directly, but you take our point.

Doing what Red Hat used to do with consumer doesn't work from the company's point of view, and doing what it's doing now doesn't work in the sense that you look at it and think, 'ah yes, a solution. It's about controlling and maybe starting to cover costs, but there's nothing there that says to you overthrow Redmond, exponential sales growth, buckets of money all round.

Nor, is there any likelihood that either Red Hat or one of the other distributions will come up with such a plan in the short term. It'll be a couple more years before a serious push for consumer is viable, says de Visser, so really you'd reckon it's a case of Red Hat et al waiting, and hoping in the interim Microsoft will lose the war, rather than any of them betting the company on actively winning it.

Nothing wrong with that, and here we'll infer for the last time today. Red Hat and SuSE in particular are unlikely to make serious plays for market sectors unless they can see how to make money out of them. Again, nothing wrong with that, but in execution such approaches grind up against the whole ethos of open source, and result in the perpetrator being accused of being 'just like Microsoft.' It seems to us that this tension cannot be resolved under one roof - the more a distribution concentrates on the bottom line, the more its free roots look like an unsatisfactory fig leaf, and the more the hard liners turn to the refusenik class of distribution.

So, what if the operations who (optimism mode on) are at least going to make money accept this, view their free operations into charitable activities and concentrate on commercial software? By making money they'll be better positioned to fund development, and as that development (so long as they don't cheat) goes back into the central pot, open source as a whole benefits. To an extent, this is what they're doing already, but an overt acceptance of 'we sell stuff, and that's cool, they do free stuff, we live together, that's cool too' would maybe help them move forward without too many people shouting 'Microsoft!' at them.

On which subject, we took de Visser to task for the clause in the Advanced Server licence agreement where Red Hat reserves the right to audit you. Sound like anybody you've heard of, we asked? The problem it's intended to deal with is clear, its effectiveness in dealing with it considerably less so, and the collateral damage it could cause to the company's image is substantial.

Red Hat has some major customers for Advanced Server who run rather more machines than they've actually licensed, and if one of them breaks - well, good heavens - it always turns out to be one of the ones that is licensed. One such customer springs to de Visser's mind (we didn't ask for names), but although Red Hat has given itself the legal right to conduct a BSA-style dawn raid, even though it's quite convinced this customer is in breach of the licence, it hasn't done so. So far, Red Hat hasn't audited anybody; in the case of this particular customer it just pulls the licence for the offending machine, and then the customer comes back with a new 'broken' one instead. Doesn't work, does it?

We suspect this particular clause in the licence won't get many - if any - outings, and as it's obviously (in our opinion) a step too far, what's it doing there? Considering the collateral damage, lame noises about the lawyers wanting it and covering yourself don't really cut it. But here's a funny thing about turning into Microsoft. If you actually make money out of selling software, then it's quite possibly difficult not to turn into Microsoft. If you're giving it away and not making money, then you don't need to be suspicious of your customers, but if you do make money, then you need to keep a lid on the pilfering. But, if you're not going to take the audit, total control-freak route - how? ®