SaaS (Software as a Service) has taken over the world. According to Research and Markets, the global software as a service market was valued at about $134 billion in 2018 and is expected to grow to $220 billion by 2022; defined as software solutions I can purchase on a subscription, or pay on a per use basis, to use an application for organisational purposes that my users access over the Web, this is probably the standard default, now, for corporate IT app consumption.
CIOs took some persuading to get here; it really wasn’t that long ago that we were being bombarded by press articles and conferences that spoke about the lack of security and robustness of this funny ‘cloud’ thing, and some national jurisdictions still don’t like the idea of data being hosted outside the country it was sourced, and so on. But, I can’t see anyone rejecting the claim in 2021 that most enterprise software is now both procured and consumed this way.
Here’s the problem: customers love the pricing approach of SaaS, the fact the job of maintaining the app and upgrading the hardware is now the cloud company’s problem, and how software is now a pay-as-you-go idea – at least in principle.
Saas, though, only really works for a very rigid class of applications. And once you step out of that box, it’sreally not that useful a way of working with vendors trying to do something a bit more complex with you—especially in the field of Artificial Intelligence (AI) and Machine Learning.
That’s a big challenge for someone else, I admit—the vendor. More and more challenges for AI companies in terms of how they operate are coming through; one of the best trackers of all this, VCs Andreessen Horowitz, has spoken very eloquently about the pressures AI CEOs find themselves having to deal with in terms of margin and scaling of their companies.
You would be entitled to say, Fine—as that sardonic Eastern European saying has it, Not my circus; not my monkeys! The corporate IT buyer has been dealing with all sorts of tricks and shenanigans from the supplier base since the start of commercial data processing, from vendor lock-in to rapacious licensing. Cloud’s ended that, surely—why should I be bothered about a software supplier’s problems making money off the model?
From SaaS to VaaS?
Don’t get me wrong, it’s an understandable position. The snag is that not being flexible could deprive you of the opportunity to really maximise your potential returns from working with AI. It could also stymie the growth of the next generation of innovative software makers, but for now let’s say that isn’t your direct concern.
Therefore, I am going to disturb your SaaS complacency by saying that to make the most of AI, you need to move from a SaaS model to what we might call VaaS—Value as a Service, so a value-based model of procuring and paying for AI solutions.
Let’s start this journey by reminding ourselves what we actually get from SaaS right now. If you think about the software as a service that’s been delivered over the last decade through cloud and the mechanisms around cloud, most of the software that’s been delivered has actually been fairly simplistic. How is it not, at the end of the day, a user interface to manage data, whether that’s your CRM or ERP data – it’s typically just managing business data in a very simplistic way.
A richer set of outputs and payback that adds value to your business
But what AI is doing for us is not just managing data. It’s about squeezing value and insights from that data. In my mind, that’s a very different proposition in terms of what needs to be done under the hood by the company delivering such a service. Rather than just building an application that’s got some nice user forms that you can fill in, manage, edit and update, and where you might have a little bit of reporting that shows how many new customers we’ve gained this month (which of course is useful) this is much more ambitious. Your AI partner is looking to explore your data to give you actual insights; these are the five customers you should be talking to this week, or this customer just received a big contract that means you could upsell.
You’re prepared to engage with AI and ML for just this richer set of outputs and payback that adds value to your business. And like it or not, the pricing model needs to reflect the value from all this. A lot of AI startups are falling over because they’re just following the SaaS model and neither they nor you as the enterprise customer are being realistic about measuring the true value that is being surfaced, and so everyone’s underpricing what’s being delivered.
Step back and look at your new AI partner again. All that extra stuff that your normal software company doesn’t have to do – doesn’t it remind you of a services company approach—for which you have always been prepared to pay something of a premium for? Service companies aren’t’t at all interested in SaaS style delivery, after all. In consulting, when you first work with a customer, you probably do start with very basic time and materials and a SaaS model, but over time, as you build the relationship, you both work to value-based pricing where both sides benefit. Without a doubt, the most attractive, but higher risk, for the consultancy or vendor, is value-based pricing where they take more risk, because if they don’t provide you with the insights that you’re looking for, you pay them less… but if they do, if they provide you with real value that you then turn into an economic profit with your customers, it’s accepted everyone gets a share in that success.
That’s actually a positive story from a CIO perspective, because it makes the vendors more accountable and more responsible for ensuring that their underlying solution is providing the gold dust that you can leverage. VaaS is a win-win for everyone in the relationship, because it means the vendors releasing more value for what they’re doing, but their business values are better aligned with the business’s values; it’s a much more positive, constructive and mature relationship.
Think about the added value that these analytics and insights you’re getting’
There’s a message for both sides of the table here from this insight. AI companies need to break away from that standard SaaS pricing model and look at how they can explain to their customers that they’re adding far more value than a standard piece of BI software. There needs to be a way to gain a share in part of that added value from the customer. And for the CIO, there’s lots of small AI vendors beavering away on very specific use cases, so if you don’t change the way you value what they’re doing there’s a risk that a lot of these companies will just fade away and disappear. So to ensure a long term and improving quality of service from these vendors, you need to think about the added value that these analytics and insights you’re getting.
You should also keep an eye on wider industry trends in terms of consolidation and how these companies themselves might want to ally with the big consultancies, which would take the decision out of your hands, I guess. It’s like with Big Data; 10, 15 years ago, you had lots of big data providers—and then eventually you’ve just got a handful building the underlying core tech that the market as a whole benefits from.
In conclusion, then: Salesforce has its famous logo of ‘No Software.’ I am not going so far as to say ‘No SaaS’—but I am definitely prepared to say, ’No SaaS AI Can Help You.’
It’s time to look again at how you buy AI if you’re serious about using it in anger.