Will composable ERP take on a different meaning in 2024?

A couple of years back, I was ensnared by Gartner coining 'Composable ERP'.

As a principle, it seemed to capture what we were seeing and hearing across a confused customer base. As a phrase, I could have chosen, better, more metaphorical terms, but in principle composable ERP was spot on.

As I've rambled previously, the 1990s and 2000s saw large enterprises implementing a new form of fully integrated business systems where everything (in terms of business processes) lived in the same box and, well, integrated.

Move some stock, a financial post happens. Take an order, an inventory and credit check happen.

This was a game changer for large complex businesses and led to huge efficiencies and visibility across their complex operations.

Vendors like SAP then bolted on ancillary (remember New Dimensions?) products that they'd either acquired or built to beat-off competition in certain areas.

The upshot, come the 2008 economic downturn was a kind of on-premise, fully integrated system with bits bolted on to solve certain new problems. Kind of like a 20 year-old new build house with a conservatory and loft conversion - it worked but was a little tired and ugly.

The problem with integrated ERP though was that changing one thing meant impacting another. So, if nimble areas of the business like Supply Chain wanted to make changes to their business model, finance and procurement had to go along for the ride.

Upgrades and wide-scale changes meant testing everything to be confident the system worked. ERP became expensive and, linked to Gartner's Pace Layering principle, challenged leadership who needed to de-couple the more strategic fast-paced business layers from the more stable governance layers.

Monolithic was (and still is) a very relevant phrase. Glacial pace of change similarly fits really well.

Cloud came along and SAP (alongside other vendors) reshuffled their hand to deal customers a new solution, new database and new hosting model - plus a new licensing principle.

The problem with the hand they dealt was that there were a few jokers in the deck.

Use Supplier Relationship Management (SRM)? Well, you now need this cloud based solution called Ariba bolting on.

Use Payroll? Well, you now need this cloud based HR solution called Success Factors bolting on.

Use our proprietary integration middleware solution PI/PO? Well, you need to find something different as we don't like that anymore.

Dig deeper and you find that lots of fragmented bits of functionality are either quite different or missing altogether.

Imagine investing in a Porsche 911 as your retirement present to yourself to be told by the German car manufacturer 5 years in that they don't sell oil filters, tyres or windscreen wipers for it anymore. But you can probably get some off a Transit or Bedford Rascal and ask a mechanic to make them fit. Alles Klar?

You'd be livid, right?

The upshot is that businesses have to make a big decision to either buy into all of SAP's non-integrated solutions (remember integration was the big upside in 2000) and do it in a certain order so that they're not impacting every business area in one shot. Or, look at alternative cloud based solutions for certain areas and then wire everything together themselves.

This is the principle of Composable ERP - you compose a previously integrated solution out of Lego blocks (or Legos to my American friends) and hope that the house you've built isn't a house of cards.

This market dynamic and shift to cloud had created windows of opportunity for niche software vendors who offer so-called Industry Cloud solutions. These solutions are designed not to solve a generic problem for a large addressable market, but to solve a specific problem for a tightly defined addressable market.

According to Gartner's Strategic Technology Trends for 2024:

By 2027, Gartner predicts more than 70% of enterprises will use industry cloud platforms (ICPs) to accelerate their business initiatives, up from less than 15% in 2023

So, S/4 HANA, Workday for HR, Coupa for Procurement and Kraken for Billing?

ERP roadmaps and product selection are now dealt out by a croupier sat in the corner with a pack of playing cards, 2 packs of Top Trumps, a pack of Happy Families and some Tarot cards.

We'll have a Queen of Spades, 8 of Diamonds, a Toyota Land Cruiser, a Stegosaurus, Mr Bun the Baker and Death please.

At Resulting IT we decided 5 years back to focus on independent roadmapping, product selection and Systems Integrator selection and bring a Sheriff's badge to the wild west. Customers massively appreciate the independent perspective and the fact we don't back-cast from a pre-determined answer to contrive a roadmap.

A couple of other back-of-the-wardrobe observations from the past few years:

1. With composable ERP, you'll probably wind up with different SIs doing different bits - and a mish-mash of methodologies.
2. Nobody will take responsibility for making sure it all hangs together except you (or us if you ask us to help).
3. Integration, Data and Analytics are most likely to fall between the slats. You'll need to have a clear architectural model and in-house skills for this stuff.
4. Although software vendors offer solutions for AI, Integration, Low Code and such like, you don't have to use their stuff and probably shouldn't if your ERP stack is heterogeneous. Even if it's largely homogeneous, there are lower cost and better alternatives to these areas - either from Hyperscalers or niche providers.

A dog has 4 legs, but not everything with 4 legs is a dog.

But, none of what I've written above by the time my large Americano dregs are cold is why I started this ramble.

No, here's my payload for you to think about...

Generative AI is the new kid on the block.

Many commentators are saying that Gen AI is going be a bigger change to our world of work than the Industrial Revolution (1, 2, 3 or 4.0) - and that it'll happen more quickly than any previous revolution did.

It's easy to think of Gen AI as creating photorealistic pictures of Elephants riding bikes, or Deep Faking politicians making false political claims (The BBC's latest Reith Lecture on this is great BTW).

And, of course, it will do this stuff and more.

But what if my croupier were to include Gen AI in their deck of cards?

What if I were to prompt ChatGPT with:

We spend ￡3m per year on this software package with a cloud-based subscription business model. Please write me a bespoke ERP solution that runs on AWS using Open Source language XYZ and Database ABC...

This ask would be unthinkable with a team of in-house developers - replicating what a large software company has spent 20 years perfecting would a ludicrous undertaking, right?

But with Generative AI?

I think the future of composable ERP will involve using Gen AI to actually compose new ERP solutions? As in build them. From scratch.

I'll leave you with this as to why it's not as ludicrous as you might think...

Standard software packages that come with best practices mean that every company who implements them is standardising on the same core business processes. They invest tens of millions of whatever currency they prefer to use to implement systems that fundamentally can't create competitive advantage.

Standard best practice processes for AZ and GSK, Coke and Pepsi, Ford and VW = zero competitive advantage.

A high-stakes zero-sum game.

Their competitive advantage from a tech perspective lies in the clever things they've built for themselves to fill white space - their customer service, product innovation, innovation management or manufacturing execution systems.

What if the future of ERP lies in every business process driving competitive advantage?

Tesla has built its own ERP solution for exactly this reason.

What if, with Gen AI and an unlimited team of developers that can work iteratively to create solutions, the disruptors (first) shun the overhead of off-the-shelf ERP and say

"Fuck it, we're building our own ERP with an army of robots?"

Asking for a load of friends.