Five questions retailers should ask before their next ERP project
Pearl hosted a webinar 18th of September built around one question: before you commit to an ERP project, are you asking the right things first? Elin Sterner and Johan Linde walked through five questions that decide whether a company is actually ready, drawn from 20 years each of retail, wholesale and manufacturing implementations. Here's the session.
1. Can today's landscape support your next business model?
“The real test isn't whether your ERP works today. It's whether it works after you add complexity,” says Johan Linde. New entities and new markets multiply complexity, not just transaction volume. Data model flexibility matters too: new revenue models expose whether your products and pricing structures were built to bend. And how your APIs were built shapes what comes next. “API-first systems let you plug in new integrations quickly. API-added systems turn every integration into a custom project,” Linde says.
A strong platform can still stall if the decisions around it move too slowly. “A great platform with slow decision makers still makes it a bad platform,” he says. Staffing carries the same weight. “A powerful system nobody can hire for creates key person risk instead of system risk.”
The test he and Elin ask every customer to run: “We call this the onboarding test. How many weeks does it take you to add a new company, a store, a country, or a sales channel? And is that number improving? If you can't answer that quickly, that's your answer on readiness.”
Elin Sterner adds what she's seeing on the ground. More companies now serve both B2B and B2C at once. Retailers want to test new ideas and channels faster, without starting a large IT project every time. Own brands mean more tracking of batches and serial numbers, and the product's life doesn't end at the sale. Repair and resale are showing up more often too.
2. Does ERP cause the problem, or just reveal it?
“Is your ERP the bottleneck, or is it the flow between ERP, POS, commerce and data?” asks Linde. In his experience, the same issues surface with every new customer. “The symptom isn't the cause. ERP takes the blame, but the fault usually sits in the integration layer, the master data, or the process behind it.”
Commerce platforms ship weekly, ERP maybe twice a year, so the two have to be treated as one value chain, not two silos. And often nobody owns the full picture of price and stock data across systems. “The customer should never notice when one system ends and another begins,” he says.
Elin Sterner sees the same pattern from the other side. “I've never been in a go-live where the customer says, we have great master data. You're very often all the same. We find duplicates, missing information, old products, and sometimes dummy articles that are still being used.” Her conclusion is blunt: “A new ERP doesn't fix bad data. You simply move the bad data into a new system.”
Ownership is the other half of the problem. “For me, there are two very important things: master data and ownership. Understand where the problem sits before you place an ERP.”
3. Where does standardization create speed, and where does it cost you?
Linde's rule of thumb: “Standardize what the customer never sees, and protect what they choose you for.” Order logic, tax engines and master data integrations are worth locking down, because none of it is visible to the customer. Checkout, payment and fraud rules are worth standardizing hard too; a consolidated setup lowers risk and pays off. Point of sale and inventory sync should be identical everywhere, while service model and layout stay local.
Elin Sterner frames it from the customer conversation she has constantly. “Standardize where standard is good enough,” things like tax, accounting, basic order processing, warehouse and transportation management rarely give a retailer any real edge. But assortment, customer experience, pricing and private label are usually where the differentiation actually lives. “Standardize where the customer doesn't care, differentiate where they do.”
She's just as direct about people. “If you don't let your people make decisions, and have room to make mistakes that you can correct and move on from, then you will fail.” And not every system needs replacing. Sometimes integration is the better first step.
4. Which AI foundation matters now?
“Every use case is only as good as the data behind it,” says Linde, which is why data strategy and cleanup have to come first. Standalone AI dashboards tend to get abandoned, so AI works better embedded into the workflows people already use. Vendor-embedded AI, SAP's Joule for example, lowers integration risk, but still has to be measured against real use cases rather than just switched on.
The number he shares stops most rooms: “95% of AI projects never reach production. A study from MIT last year found that 60% of enterprises evaluate AI, 20% get to a pilot, and only 5% reach production with a measurable profit and loss impact.” His conclusion is that governance has to come before adoption. “Decide who approves AI touching money, customer data, or any other critical data before it's live, not after.”
Elin Sterner describes what embedded AI should feel like day to day. “If I work with a sales order, I want AI to help me where I'm already working. I don't want to copy information into another AI application.” That's the idea behind Joule. But the ceiling is still data. “If your product data is wrong, customer data is duplicated, or stock information is unreliable, AI will not magically solve that problem.”
Linde's advice on where to start: “Don't start with how many AI agents can be built. Start with where AI can create real business value.”
5. What should you do now, and over the next five years?
Linde lists six things worth doing before any ERP replacement starts: treat master data as a strategic asset and clean it now, since “no new system fixes bad data, it only inherits it;” inventory your customizations and retire the ones that no longer earn their place; decouple POS, commerce and ERP through APIs so integrations don't become rework later; build a financial case by tracking the real cost of today's limitations; skill up on what's coming; and choose your implementation partner early.
His closing point: “The ERP migration doesn't start when the contract is signed for a new ERP solution. It starts with the decisions you're making about this year's data.”
Elin Sterner's version of the same advice, for anyone whose ERP replacement is still three to five years out: “Measure onboarding. Ask your IT team, if we buy a company tomorrow, how many weeks does it take to bring it into your system, and is this number going down or up?” Think about traceability too, especially with your own brands, manufacturing, repair and resale. And start from the business, not the system. “Ask what we want to be better at in the future. What information are we missing? And what should be automated?”
Three questions to take with you
How many weeks does it take to add a new company, a store, a country or a sales channel, and is that number improving? Do you manage price and stock data across your systems successfully today? And if AI had to run and combine your ERP, POS, commerce and other system data tomorrow, would you and your customers trust it?
None of these start with picking a system. They're about how ready you are before you do.
If you want to go deeper on your own retail landscape, reach out directly to Elin or Johan.
