In Andreessen’s original formulation, PMF was binary. There were BPMF and APMF: “Before Product Market Fit” and “After Product-Market Fit.” In that model, companies iterate until they find PMF, then scale. Graphed, it would look something like the following (imagining sales as the Y-Axis):
While this might fit a Facebook or a Google, most B2B markets are more complex than this. Business-to-business markets are usually a hodgepodge of multiple types of customers, product use cases, purchasing processes, and geographic eccentricities. In these markets, getting to initial PMF is not that big a breakthrough – it’s usually just the start of a long process. As the company scales in these markets, it has to continually re-create new product market fits as it reaches saturation with the customers and use cases it can reach with existing products, channels, and well-served geographies.
So, rather than the “no PMF – PMF! – Scaling!” model, it’s a better idea to expect a model that looks more like the graph below: a sequence of new channels, products, and geographies that together add up to a smoothly scaling revenue line. A company that successfully navigates this process is achieving more than just PMF: it’s achieving Go To Market Scale.
In the real world, once a product achieves its initial product market fit, it attracts new competitors almost instantly, saturates the easy channels quickly and usually has to add capabilities to serve new geographies.
While the simple model of Product Market Fit can help focus teams that are trying to get someone (anyone!) to buy their first product, I think it leads teams to think “job’s done” when they hit on the first working combination of product/channel/customer. After initial PMF, many teams can ramp quickly to their first $3M (or $10M or $40M) of sales and then get blindsided when growth slows. And it’s understandable why this happens: it’s hard to admit that things need to change, particularly when the team went through multiple painful iterations to get to their initial PMF. But too often, teams react by doubling down on the current go-to-market mix through increased sales hiring or bigger marketing expenditure even as the incremental sales head or the new marketing spend returns increasingly disappointing results. The next step is usually to fire the sales or marketing head for poor execution. But it’s usually not the execution that’s at issue, it’s the go-to market (GTM) mix.
CEO’s and their teams should continually reassess and evolve their go to market mix to keep growth on a smooth ramp.
It’s not a sign of failure when a particular GTM mix starts to yield diminishing returns. It’s normal for channels, products and use cases to saturate. It’s only a failure if the team doesn’t detect it and react in time to keep sales ramping.
Go To Market Scaling works by executing a smooth sequence of individual product market fits, each time establishing:
• Value Fit: The product delivers significant net value for a use case
• Purchaser Fit: There is an identified organizational purchaser
• Distribution Fit: There is a cost-effective distribution channel
Most teams focus on the first two bullet points: making sure that the product is delivering sufficient net value for well-identified use cases. This is where tools like Net Promoter Scores and “MustHaveIt” scores help. Scoring high on these measures means that products are delivering value. But if it’s tough to get customers to be references or if your product reviews are less than enthusiastic, then the company doesn’t have Value-fit (at least for these customers and use cases).
But, the third bullet point above is equally as important and often ignored: how do you know what channels are likely to be a good distribution fit and how do you know when you have good fit? The latter question is usually easy to answer: when you have distribution fit it’s suddenly easy to recruit new high-quality channel partners or top performing sales reps from other companies. Almost everyone makes quota. Highly qualified prospects complain publicly that they can’t get a call back from your sales force because you don’t have enough sales capacity to handle demand. Alternatively, in a self-service SaaS model, your word of mouth driven direct traffic or your referral/viral loop suddenly starts kicking in. Everyone wants in.
Once it happens, you can see that you have fit – but it’s a harder problem to figure out what distribution is going to be the right fit before you go to market. Part of it is understanding your unit of value. But it’s equally as important to understand your product’s selling costs, first unit implementation cost and how your products’ value scales with units.
The value that a product can deliver when it’s an Enterprise standard vs. a team standard vs. single user is almost always qualitatively different.
If the cost of adopting the very first unit is high (perhaps because of a security review or data integration requirements), then those costs put a minimum bar on the number of units you have to sell at one time – which in turn limits your choice of distribution.
If you understand your product’s value scaling and minimum adoption cost, you’ll be on the way to picking distribution with good fit. For example, products with high implementation costs usually require a direct sales force equipped with systems engineers and professional services.
In Marc Andreessen’s original blog post on Product Market Fit, he mentions VMware as a product that was so revolutionary that it skipped right past product-market fit: “bring a product as transformative as VMWare’s to market and you’re going to succeed, full stop.” While encouraging to read, those of us who were at VMware in the early 2000’s definitely did not feel success was inevitable. When I joined VMware as a product management director in early 2002, the dominant view of the company in the valley was that VMware was an interesting $20M dev tools company struggling to break out into broader relevance.
In early 2002, VMware had been selling two server products for a year: GSX Server and ESX Server. GSX Server sold through VMware’s existing channels for its desktop Workstation product (website and volume resellers/catalogs) and had successfully generated a few million in sales its first year. This was a solid success for a company with mid-20’s million in sales. On the other hand, ESX Server – VMware’s more strategic server product – went to market via a traditional Enterprise direct sales model (4 account teams composed of a direct rep, systems engineer and inside sales rep). And in contrast to GSX Server, ESX Server sales had disappointed. In that first year, the account teams had only sold $350,000 worth of product to a total of 70 customers. While there were several large Enterprises like GE Medical, GE Industrial and Merrill Lynch in proof of concept projects – sales velocity was poor. Because the company had chosen an Enterprise direct sales model, each sales team had to sell large numbers of units to each account in order to cover sales costs and make the model work. But at that early product stage (1.0), the product was too immature to run production workloads, there was no management product to help customers manage large numbers of virtual machines and the product wasn’t ready for data center environments with demanding SAN, backup and DR requirements. As a result, the only large Enterprise ($1M+) that eventually converted was Merrill Lynch.
But Diane Greene, the VMware CEO, quickly realized that this strategy wasn’t working and she brought in a new sales head who quickly changed distribution strategy to drive smaller deals as an add-on in-progress server deals by IBM and HP channel partners for server consolidation projects. The change in distribution strategy immediately paid off in sales momentum the next quarter. Coupled with more focused product positioning, better sales enablement, and improved product maturity, (including a stellar management product – Virtual Center), VMware had the Go-To-Market mix that would bring it to its first $1B in sales.
The Enterprise Direct model wasn’t a priori a bad idea. My sense is that most companies would have just gutted it out through this stage until the product was ready to be sold through a direct sales force (about 18 months later). But this would have put the company at serious risk. This was not the only GTM change that VMware navigated successfully. In the late 2000’s the company successfully introduced additional management products and re-transitioned to a hybrid direct sales model with Enterprise Licenses and price segmentation that successfully continued the company’s ramp to its current $6B+ run rate.
In the VMware case, the answer to its distribution challenge was to move to smaller deals through cheaper channels sold to the specific use case of server consolidation. But illustrating the opposite case is Delphix (an Icon portfolio company). Delphix makes a data virtualization platform that allows data consumers to quickly and easily take copies, clones and transforms of any data set for development, QA, compliance, or any other purpose. Delphix theoretically has a small unit of value: an individual DBA using Delphix for a single database gets a lot of value from using it. So initially, Delphix went to market at a low price, high volume strategy through channels and self-service. Only to find that minimum adoption costs were quite high: each new species of database configuration required special tweaks to the point that the value created was overwhelmed by implementation expense. In response, the company switched to a higher touch Enterprise direct sales model that targeted larger minimum deployments with more implementation support. This turned out to be the right distribution fit and sales subsequently grew rapidly.
It’s a fantastic feeling for a company to achieve initial product-market fit. But for most B2B companies, this is just the start. Seeing the early signs of saturation in current Go To Market mix, and reacting quickly is a vital skill for every growth company to learn.