The topic of pre-chasm startups is one I’ve long been obsessed with, as a former entrepreneur who (along with many colleagues!) founded a company around a pre-chasm technology (software-defined networking), and was involved from academic paper to product line. So it’s no surprise I’m particularly interested in enterprise startups that haven’t yet crossed the chasm (hence “pre-chasm”), even recently arguing that it’s ok if some never do. Because the application of the crossing-the-chasm model is not always straightforward in practice. This is particularly true as trends in open source, cloud, and bottom-up services drive technology adoption within the enterprise.
Geoffrey Moore — who first published Crossing the Chasm in 1991 — wrote a wonderfully thoughtful response to my piece, overall observing that “as with all ideas that get spread virally, [the meme of crossing the chasm] has accumulated meanings that were not intended, and in some cases are just plain wrong”. I highly recommend you read the full response and also recommend the book. Andy Rachleff (lecturer at Stanford, founder of Benchmark, and one of my board members for over 5 years) quoted from it often, and our company used it as the foundational model for thinking through our go-to-market strategy. We were such fans that we asked Moore to come and speak with the execs at VMware (our acquirer) in order to further dig into the ideas and implications as we brought our new product to market; that too was hugely useful and underlied our strategy in what is now a billion-dollar product line.
So, I’m a fan of Moore’s work and agree with much of his post. But the broader point I’m making, and would like to further emphasize in this post, is that the application of the crossing-the-chasm model isn’t straightforward. From the perspective of a founder, the model doesn’t always play out as you’d expect; there is a lot of nuance to applying it in practice. And I’d argue those nuances are getting more pronounced given the faster pace of technology today and the secular trends I mention above.
Below I’ll focus on just the first two of the areas we covered in the previous posts, although similar arguments can be made throughout.
First, let’s take the point about geographic expansion. I had originally stated that geographic expansion was yet another chasm. As Moore reminds us, the three dependencies for crossing the chasm are “a compelling use case that will create pull, a whole product that nails the use case, and a word-of-mouth community that can communicate and reinforce the marketing message [italics added].”
He further adds that when you enter into a new geography, “it is very hard to maintain this discipline, particularly if you are going through a product-oriented distribution channel”. So while it’s true that the model then holds here in theory, in practice, it’s messy, and not just because it’s hard to “maintain the discipline”. It’s because in applying the model, many founders make the inferential leap that a market exists with a word-of-mouth community if the other two conditions are met.
In our case, we knew that the NY financial sector spoke frequently with the UK financial sector, and since we had a lot of positive customer traction in the former geography, we chose to go internationally along this vertical precisely because we were trying to stay within a word-of-mouth community. However, we quickly found out that in the UK, the companies we were targeting had very different competitive environments, often had different technology and operational stacks, and relied on entirely different partner relationships. Particularly relevant here is that much of the technology stack (and the surrounding ecosystem) were driven as much by organic, bottom-up adoption of open source technologies as much by top-down vendor sales, and were different enough from their NY counterparts that we couldn’t sell to them.
While we ultimately figured it out, it took years, and by then we’d already tackled multiple other unrelated verticals. If I were to do it again, I would have first focused on expanding to those different verticals within the U.S. rather than trying to expand within a reference/ word-of-mouth community internationally. For us, the chasm across geographies was far wider than the power of word-of-mouth networks that could bridge that chasm. Now, this point may seem obvious to those who have done this before. But for a first-time founder like me it wasn’t, and unfortunately, I’ve seen this go similarly sideways with a number of other companies I’ve observed over the years, too.
The other point I want to make in response to Moore’s argument about incumbents lack of desire to compete with startups:
“Large incumbents have neither the means nor the interest to block a chasm-crossing play because they cannot operate at the small scale involved. $50M to $100M markets are a godsend to a fledgling start-up, but they are a waste of time for a multi-billion-dollar enterprise. What they can do is pre-empt a tornado market or, worst case, destroy it by blocking access to the disruptors, but this ploy only works well in consumer markets. In B2B markets, as long as the offer creates a genuine pull, your customer base can pressure the incumbent to play ball.”
I disagree. The enterprise landscape has evolved to the point where not only do incumbents bother to compete with pre-chasm products, they’re even more able to do so because of cloud and open source. This is especially pronounced where incumbents own the technology distribution channel. Take for example AWS: They have full visibility on all the services that run on them; if they see a particular service gaining traction, they can easily offer the services themselves if the code is open source.
In fact, this has been such an issue that startups have taken on the potential fallout with their open source communities to change their licenses to disallow this open behavior. Confluent and MongoDB are just two recent examples; as Mongo co-founder and CTO Eliot Horowitz said: “Unfortunately, once an open source project becomes interesting, it is too easy for cloud vendors who have not developed the software to capture all of the value while contributing little back to the community.”
And often, open source projects are “interesting” long before anyone figures out how to monetize them. The examples are many: In 2010 when Rackspace announced Openstack, it changed the entire landscape for cloud orchestration, causing many startups in the space to pivot or exit soon after. At the time, on-prem cloud offerings were still “pre-chasm”, and therefore a tough sell, yet it was a sensible move for Rackspace to gain relevancy and mindshare. Google later changed the landscape again with Kubernetes and Istio, which have shaken up cloud orchestration and L7 services within the datacenter.
In every one of these instances startups in the space are having to contend with moves by an incumbent.
Furthermore, the observation that “What [the incumbents] can do is pre-empt a tornado market, or worse case, destroy it by blocking access to the disruptors, but this ploy only works well in consumer markets” is no longer true. With the organic nature of bottom-up and developer-led adoption, enterprise companies are starting to see more and more consumer-style growth curves. As such, we see incumbents release open source or free services to pre-empt a market (or destroy it before it can flourish).
I think the incumbents have learned the lessons from the history of disruption and now realize that even a non-revenue-generating developer phenomenon can still be a huge existential threat. (It certainly was the case when I ran a business unit of an incumbent.) This isn’t to say companies can’t navigate these situations; they can! I just wouldn’t base my strategy on the assumption that the incumbents will not and cannot respond. They can and they do.
I want to thank Geoffrey again for this respectful exchange; I remain a huge fan of his and of the model. But for those companies that are struggling in the chasm and finding it to be less neat than they (or their boards) expect, I hope this back-and-forth offers some solace, if not some guidance, in the even messier world we live in today.