Yes, in Model A you may have two systems to support the two businesses. In Model B you would try to “hack” your existing systems to support the new subscription business. Model C says let’s go big bang and transform the company overnight.

Now given that Model C is your eventual goal, which approach should you start with?

We’ve been working long enough together that I can confidently say my answer is the same as yours, Tien. It’s door number one: Model A, build a separate system. Now, that’s not always intuitive.

We’ve spent what feels like the last 30 years standardizing our IT architectures. Too many systems, no single source of truth, so we’re all trained to try to build one system to handle the whole business. But it turns out if your goal is to grow the new business as fast as possible, for 80% of companies Model A is the way to go. For the rest, it’s Model C, but you have to be really, really committed. And everyone should avoid Model B. That’s a path to failure.

Okay, we’re agreed. But why is Model A the clear winner?

Let me give you two reasons. First, most companies aren’t going to get it right the first time. They’re just not. Why? Because this is a whole new business model and mindset. And in Models B and C, by definition, failure is going to be systemic. It’s also going to take you longer to get to the point when you realize it will fail! In Model A, iteration and experimentation are part of the process. You’re light, you’re nimble, you can pivot.

Second, even though Model C is the end goal, most companies are not going to have the management buy-in, resources, and staying power to pull off a huge effort like that straight out of the gate (see my first point). It’s hard to keep the momentum going behind a three-to-five-year transformation effort. There’s inertia, there’s turnover. I won’t name names here, but you don’t want to be in the position where a big high-profile corporate project is still limping along, but most of the decision-makers have all left the premises, and no one is really sure why they’re doing it anymore. You’re better off using Model A as a learning platform that informs the rest of the process.

Okay. But let me play Devil’s Advocate here. You’ve already invested a lot of time, effort, and resources into your current organization: product, development, marketing, sales channels. Why can’t you just make some adjustments? Why is Model B such a non-starter?

First of all, I think the Devil should start looking for new representation. That was a pretty weak defense.

I know. I’m trying. Humor me.

As we’ve been saying since Day One, product-0riented systems are simply not built to handle customer-oriented business models. There is simply too much detritus in these legacy systems. Decades of tinkering with monolithic codebases have made them simply too complex. You can’t just tweak a few things, or think it’s about simply chopping up your product costs into monthly increments. The culture, operating processes, and business systems are completely different. Oh and by the way, chances are these legacy systems are completely product-oriented, and simply not built to handle customer-centric business models.

That being said, you can build a new growth business with the goal of eventually “bending” your legacy systems to the new platform once it starts really taking off. We’ve seen this happen plenty of times. Once your new experiment represents over half of your overall revenue, then it’s not an experiment anymore, right? It’s your business.

I know that you deal with a lot of companies that have tried Model B or Model C and failed. What are some of the reasons why?

If they’re walking away from a failed Model C scenario, they usually bit off too much, or they lost sponsorship. For example, today I had lunch with the CIO of a big industrial manufacturing company. They’re used to selling big pieces of heavy equipment. They know they need to shift. They have a big vision for services and subscriptions. But they acknowledge that they haven’t built the right competencies yet. So we’re going to work with them to get some new businesses off the ground, but more importantly, learn and figure out what they don’t know today. If they’re coming off a failed Model B experiment, they usually tried to jury-rig a pre-existing system and they couldn’t get it to work.

So assuming you make the wise choice to go with Model A. What do you need to be successful?

First, get a business sponsor that aligns with a digital-first mentality. Don’t try to teach an old dog a new trick (with all due respect to all the old dogs out there). Because when you get successful at subscriptions, you’re going to start challenging a lot of previously accepted assumptions. So your leadership is going to need the ability to move beyond those old behaviors. Second, make sure you figure out how you’re going to sell it. What is the distribution channel going to look like? What do the feet in the street look like? Because it’s great to have a new digital service offering, but if your sales channel isn’t properly incentivized to focus on it, it’s going to die on the vine.

Okay, let’s say things are going well, and moving from Model A to Model C is starting to become a real possibility. What does that transition look like? How do I show my boss that this is really working?

Well, to start, you’re seeing volume go up. I’m not even talking about dollar volume, but volume of successful capabilities launched into the market. So if you can show your boss that you launched several new services, and it took you a few months as opposed to a few years, that’s going to show growth-oriented leaders that they have a new way of doing business. And in addition to volume, you need to look at growth rate. Say you’re working at a huge company that’s growing 5% a year. If you can build half a dozen small businesses that are each growing at 20% a year, then sooner or later, you’re going to get someone’s attention. You’re going to reach a tipping point.

Those are also two proof points in favor of Model A: you can move faster with discrete systems built exclusively for services, which also means that those services can gain traction quicker.

You’ve built yourself an experimentation platform. You can learn.

Thanks Robert!

Thanks Tien!