The signals it really is time to look at ERP
There are a small number of conditions where a specialist manufacturer should genuinely scope an ERP upgrade. Stock value is consistently and meaningfully wrong at month end, and the gap cannot be explained by counting frequency. Production scheduling is run from spreadsheets that one person owns, and that person is now the bottleneck. The shop floor has moved on, but the system still describes the business as it was three years ago.
When two or three of those signals are true at once, the conversation is real. The current system is no longer a record of the operation, it is a daily source of friction that costs management time, customer trust, and cash.
The much more common signals it is not
Every other week we meet a manufacturer who tells us they need a new ERP. When we look closely, the system is not the problem. The problem is that nobody owns the part master, two teams use different units of measure for the same item, and the BOM has not been reviewed since the last product family was introduced. A new ERP will not solve any of that. It will inherit it, magnify it, and the project will be remembered as a failure.
- Sales orders are being keyed twice because two departments do not trust each other's data.
- There are three live spreadsheets covering capacity, and none of them agree.
- The MRP run is switched off because it generates output nobody believes.
- Operations management has stopped reading the standard reports.
The four-category UK shortlist
When the scope is real, the shortlist for a UK specialist manufacturer almost always falls into four categories. We have implemented or worked alongside all four.
- Sage 200 for businesses that grew up on Sage 50 and want a familiar finance core with light manufacturing. Cost-effective, well-supported in the UK partner channel.
- Microsoft Dynamics 365 Business Central for businesses that want native integration with the Microsoft stack (Teams, Power BI, Power Automate) and a stronger manufacturing module than Sage 200.
- NetSuite for businesses that have outgrown SME-tier products, run multi-entity or multi-currency, and want a single global system with proper API depth.
- A specialist MES layer such as Tulip or Katana sitting alongside a smaller finance ERP, for businesses where the shop floor is the bottleneck and the finance system is fine.
What to fix before the scope conversation starts
The work that actually determines ERP success happens before anyone demos a product. We insist on five things being in place before we let a client open an RFP.
- A single owned part master with clean units of measure, costs, and lead times.
- Reviewed and current BOMs for the live product range, with a documented update process.
- An agreed inventory operating model (where is stock counted, by whom, on what cadence, against what tolerance).
- A named process owner for sales-to-cash and procure-to-pay, with mandate.
- A short list of the five reports the business actually runs the operation from.
How to scope so the ERP does not eat 18 months of management bandwidth
The default ERP project absorbs every senior person in the business for a year and a half. We do not run them that way. The scope we recommend is narrow first, broad later: get finance, stock, sales orders, purchase orders, and basic production live in a thin first phase, then layer in scheduling, quality, MES integration, and reporting once the operation is stable on the new core.
This is unfashionable. Most partners want to sell a complete transformation because that is what their pricing model rewards. It is also the single biggest reason ERP projects overrun. Gartner's ongoing ERP research consistently finds scope expansion to be the most common driver of cost overrun and the most likely cause of go-live being delayed.
When a phased approach beats a single cutover
If the business has more than two manufacturing sites, or more than one legal entity, or a meaningful seasonal peak that cannot be moved, a single cutover is rarely the right answer. We default to a phased approach in those cases: prove the new system on one site or one entity for one full month-end cycle, fix what surfaces, then roll out to the rest.
The phased route looks slower on paper. In practice it is almost always faster to a stable, trusted system, because the team learns on a smaller surface and the production operation never goes dark.
If you are weighing this decision now, our short answer is the same one we would give a client on a call: do not start with the product, start with the part master. If the part master is clean and owned, almost any of the four shortlist categories will work. If it is not, none of them will.