Positioning and Branding

by Tim O'Reilly
July 1997

This month, I thought I'd share with you some of the marketing concepts that have been increasingly shaping my thinking about our successes and failures over the past years. This may seem out of the blue, but may give some interesting perspective on how things hang together in my mind.

The first concept is from books by Al Ries and Jack Trout, including Positioning and The 22 Immutable Laws of Marketing. Ries and Trout argue that you need to own a consistent position in the mind of your market (defined by Geoffrey Moore, who I'll talk about in more detail later, as a group of customers who refer to each other when making buying decisions.) Only one company can own a particular position; anyone else trying to own that position is an also-ran. Positions can evolve over time, as companies try to come up with a position that is more compelling than someone else's existing position. Ries and Trout express this point by saying: "If you can't be first, create a new category in which you can be first."

For example, there was a time when Hagen-Dasz established a new market position: "premium ice cream." They created a European-sounding name (despite being a product of someone like Pillsbury), and used rich ingredients and taste, a high price, and pint carton (vs. the quart and half gallon sizes that dominated the market at that time) to establish their position. Once they succeeded, a number of other competitors tried to occupy the same position, but none of them did as well...until Ben & Jerry's came along, and established a new, more specific position: the premium ice cream for environmentally friendly, socially conscious consumers.

If you look back at our success with computer books, it is clear that we've been able to establish a very strong position, which has evolved from "the leading publisher of books on UNIX" to "the leading publisher of books for serious computer users."

Any time we've done books that have strayed from that position, our books haven't done anywhere near as well. Good examples are the What You Need to Know books and the Songline Guides. Both are very well done, interesting books. But they don't appeal to the same market or share the same position.

You can see from these examples just what "market" and "position" mean. For example: how can you say that "What You Need to Know When You Can't Find Your System Administrator" doesn't appeal to the same market? It's about UNIX--in fact, it's about UNIX administration--a core topic for us. But considered in light of our positioning, it's worlds apart. This book (and in fact this series) was aimed at casual computer users, people who don't really want to be using computers but have to for one reason or another. That's the position that IDG has exploited so well with its Dummies series. It's a mistake to equate "market" with "topic" (although there are obviously a lot of ways that the two overlap and influence each other.)

By contrast, the In a Nutshell series, as we now envision it, will be a broad-based series, covering topics that may stray far from our established subjects (Photoshop is a good example), but I will venture to predict that they will be more successful, because they stand on our core position: these are books for serious computer users. The catch, as we get farther and farther from audiences that know us, will be making sure that our position is known to our market (those customers who refer to each other). In the retail channel, this is pretty clear: the retailers understand our positioning, and our extension to other topics with the same kind of high end books makes sense to them.

Timing may be one important success factor in extending our position into new markets. My hope was to have Win95 in a Nutshell be the lead in the brand extension, but it was delayed by an author who didn't come through. (It's now scheduled for January.) Still, I think our brand (which is the visible embodiment of our position) has gotten well enough known that we will carry off the broader In a Nutshell launch anyway.

While I'm on the subject of branding, I want to note that we've established a very powerful brand with Edie's animals. One of the things that is extremely important about a brand is its consistency. So, despite pleas from booksellers and customers, we've refused to put animals on books that seemed to us to have a different character than our core books. But as Ries and Trout note: "Sacrifice is the essence of positioning." You can't be all things to all people. You've got to decide what you're going to be, and stick with it. Similarly, Edie has continually argued for the purity of the images in our marketing--that we shouldn't doll them up with cute gimmicks or elaborate them in too many ways. If you think about great brands, from Marlboro (yes--a great brand if a bad product) to Coca Cola, they tend to be restrained and repetitive--perhaps insistent would be a better word--in the use of their brand image. The Marlboro cowboy is the longest running ad campaign in the history of modern advertising, and arguably the most successful. It hasn't changed much in 30-40 years.

In short, as long as a position and a brand are working, keep going with them. As we have experimented with various looks that depart from our core, well-known brands, we've learned (sometimes the hard way) that while it may not be as interesting as doing something new all the time, reinforcing the brand and the position with each new product and each new marketing piece is very important.

That doesn't mean that you can't innovate. You have to, despite the risks. And in fact, if you look at how our books have evolved, we've changed them quite a bit. But the most successful changes have been in the direction of clarifying the positioning and the relationship to the brand.

Again, the In a Nutshell books are a good example. When we first released UNIX in a Nutshell, we didn't put an animal on it, because it was a different kind of book from the regular animal books, and we didn't want to confuse the market. After Java in a Nutshell became a big hit, we somewhat reluctantly (with some prodding from retail and a marketing focus group) agreed to use a variant of the animal design on the In a Nutshell quickrefs.

What's interesting is that this decision really helped to clarify in my mind the editorial positioning for the series as well. The books had to be high end, because that's what the animals mean to our customers. It became clearer and clearer to me that rather than being a dilution of our position, using the animals on the quickrefs could be used to reinforce it. Accordingly, I've crafted a whole story (which I repeat to press and booksellers) that there is a big change in the computer book market, with more and more serious computer users--people who use computers as an important part of their job, people who've used computers for years, and don't need a lot of handholding, but just the facts.

Let me jump sideways to software. When we first released WebSite, we used the positioning: "the web server for everyone." It seemed like a good idea at the time. We did (and still do) believe that "everyone who has a web browser ought to have a web server." These were good slogans, but they had two problems:

1. They didn't trade on or reinforce O'Reilly's existing positioning in the market.
2. They described a market that was too large for us to capture successfully.

This latter point is driven home by William Davidow, whose book Marketing High Technology makes the point that the size of the market you can successfully dominate has a fairly strong mathematical relationship to the size of your company and its resources. Davidow argues that in order to create a self-referencing positive feedback loop in a market, you need to have dominant market share--ideally more than 50%. So that means that if you want to succeed in a billion dollar market, for example, you need to be, or be able to quickly become, a $500 million dollar company.

This might seem counter-intuitive. After all, look at all the companies that started small and got big--Lotus, Microsoft, etc. Well, the fact is that the markets they came to dominate were small when those companies started out in them. They grew with the market.

Similarly, look at us in computer book publishing. This is a relatively small market as a whole (a few hundred million dollars); the high end of that market is considerably smaller. We were able to achieve a dominant market position in UNIX books, and from there gradually extend our reach into other high end areas. But note what happened with the Internet. We started out dominating a small market category (internet books), but when the category exploded, we didn't keep up. There was a period in which, if we'd released more books, we could have solidified our market dominance, but instead we became one of the pack. In retrospect, we should have done more of the second generation Internet books we now have out there somewhat earlier, when we could have used them to solidify our position in the market.

Back to software. Clearly, the PC web server market was small when we started with it too. But hindsight tells us that we weren't prepared to grow as fast as the market, which meant that we would eventually be marginalized.

What do we do about this? Just throw in the towel?

In his books Crossing the Chasm and Inside the Tornado, Geoffrey Moore talks about the problem of transitioning from success with early adopters into success with mainstream markets. His insights are complementary to those of Ries, Trout, and Davidow. He argues that you need to find a foothold in a niche market that you can dominate, and from there expand into other markets.

That's what we're struggling with now. It's clear that the general purpose web server market is too big for us, with very large, deep-pocketed players battling it out. We've realized that far from being "the web server for everyone", WebSite is the web server for technically sophisticated users who can tell hands-on that it's a better product. As Bob Denny noted, since large company server purchases are often mandated by central departments, this means we're typically looking for small professional businesses, where the person making the purchase decision is likely also to be the person who may be implementing the server.

This points us to particular segments of the small business market--computer consultants, web site designers, and sophisticated small businesses (perhaps law firms, architects, accountants?) Mostly we're getting back to what we've always done with our books: targetting people like ourselves, with problems that we know how to solve.

That brings me round to Perl. One of the things that's really nice about Perl is that it's a "right-size" market for us. It's hot, but not so hot that it's going to be a billion dollar market. If Perl and Perl products became a hundred million dollar market, we could still be a dominant player.

Obviously, there's a lot of synergy between the market for Perl software products and Perl books that we can exploit (in a way that we couldn't exploit the book connection with WebSite because the software was out in front of our publishing program). With the resource kit product, there is also a retail channel we can count on (versus the boxed software channel, which imploded in 1996.) But perhaps most importantly, the Perl Resource Kits share our proven position--products for sophisticated computer users.

One of the marketing challenges I see for software in the next six months is to tie our existing products more clearly into this position. I think we'll have some great opportunities to do so by cross-fertilization between WebSite and Perl (especially with the NT Perl Resource Kit). We want WebSite to be (and to be known as) the most Perl-friendly win32 web server. In this way, all of our stories and positions start to line up, and reinforce each other rather than standing alone.

Now, this is a very high level view, and leaves out a lot. It also probably ties these ideas down to some specifics that won't make sense in the light of closer analysis. But I am quite confident about some of the bigger issues: that O'Reilly owns a position among computer professionals and other serious computer users that is an essential strong point. Both our products and our marketing should be designed to reinforce this position rather than to dilute it.

One final source I want to cite and add to the mix is an excellent paper that Eric Raymond delivered at the Linux Kongress in Wurzburg earlier this year (and will be delivering again at the Perl Conference.) Entitled The Cathedral and the Bazaar, the paper explores what makes collaborative freeware projects like Linux so successful. Eric makes many good points, but the one that I found particularly compelling in the context of this article is that free software has the characteristics of a "gift culture."

Anthropologists have recognized that under conditions of abundance, cultures evolve a situation where you get status not by what you own but by what you can afford to give away. This was the origin of the "potlatch" concept among the Indians of the Pacific NorthWest.

Eric points out that free software is a gift culture, and people do indeed get status by what they give away. The secret, then, to a successful collaborative freeware project is to create opportunities for more people to get status. He argues that the more Linus Torvalds gives credit to others for advancing Linux, the more they think he is something special. Eric says that Linus has managed to create "an efficient market in status."

This is a last piece of the positioning mix that I'm trying to understand. There's something to the whole gift culture thing. But even more, there's something to the idea of an efficient market in status.

This seems to me to be an incredibly interesting thing to think about. First, and perhaps most trivially, O'Reilly books have definitely become a status thing. You get status for owning them; you get even more status from writing them. Like giving away your software, creating an O'Reilly book confers status. Perhaps more importantly, taken over to our software business, what it says to me is that one of the ways to create buzz around our products is to find ways that they can help confer status on their users.

It was this point, among others, that prompted me to say "Let's do a Perl conference!" That's a great way to get users together, but also to confer status on them. When you do that, they start working to get more status...

This is exactly what is happening. People are excited to become presenters. They want to work with us more closely, whether on books or software products...

I know I've drifted from my original theme here. I can't pretend that these thoughts are fully cooked. I just wanted to share the mix of concepts that is providing fertilizer for the next round of product development and marketing ideas. Let me know if you found this interesting.