| Start by ruling systems out, not in. Manufacturing execution systems, laboratory information systems, automation and generic ERP without formulation logic will all demonstrate well against a loose brief and reveal the gap during implementation. Once the shortlist is genuinely in the category, score it on six things: whether the platform understands formulas rather than bills of materials, how its regulatory content is maintained, what happens at the two handoffs where formulation data is most often lost, what implementation actually involves, whether the vendor will still fit in five years, and what comparable customers report. Then test three scenarios live in the demonstration: a regulation changes, R&D approves a formula, and a supplier sends a specification. |
Why Do Formulation Software Evaluations Go Wrong?
| Most evaluations fail on scope, not on scoring. Adjacent systems answer a formulation brief convincingly in a demonstration and reveal the gap during implementation. Establish the category boundary before you build the shortlist. |
The common failure is not choosing the wrong vendor from a good shortlist. It is building a shortlist of systems that were never in the category.
Formulation management software manages the formula itself — ingredients, quantities, interactions, cost, and regulatory status — across the product lifecycle. Four kinds of system sit next to it and are regularly mistaken for it.
Manufacturing execution systems run production. They execute a recipe; they do not develop one.
Laboratory information systems manage samples, tests, and results. They are built around the experiment rather than the formula, and they integrate with formulation platforms rather than replacing them.
Equipment control and automation solve a different problem at a different layer of the stack.
Generic ERP without formulation logic knows a bill of materials. It cannot evaluate what substituting one ingredient does to an allergen declaration across markets, or whether that substitution breaches a cost threshold that makes the product commercially unviable.
Any of these can demonstrate well against a loosely written brief. Rule them in or out before anything gets scored.
How Do You Build the Shortlist?
| Vendors reach formulation from different starting points — process PLM, enterprise resource planning, laboratory informatics, and scientific modeling. Each brings real strengths and a predictable blind spot. Knowing which lineage a vendor comes from tells you what to probe hardest. |
Shortlists in this market rarely compare like with like, because the vendors arrived from four different directions.
Process PLM platforms were built for formulated products. Formulation, specifications, and regulatory content are native. Probe the breadth of enterprise integration and the depth of scientific modeling.
Enterprise resource planning vendors extend a system of record downward into product development. Transactional and financial data is strong. Probe whether formulation is genuinely modeled or simply stored, and ask what the system does with an ingredient substitution.
Laboratory informatics vendors arrive from the experiment. Sample, test, and result management is strong. Probe specification governance, labeling output, and supplier collaboration.
Scientific modeling and materials informatics vendors arrive from prediction. Simulation and optimization are strong. Probe regulatory coverage, market-by-market compliance, and production handoff.
A shortlist that mixes lineages without accounting for this takes longer and produces scores that are not comparable. Decide which starting point suits your problem, then compare within it.
What Should You Score a Shortlist On?
|
Six areas, ordered by how early they can end an evaluation. The first three test whether the platform does the work. The last three test whether the deployment succeeds and holds up. A platform that passes the first three and fails the last three still fails. |
Use these six as a checklist against any shortlist. They are ordered deliberately: the earlier ones disqualify faster and cost less to test.
1. Does it understand a formula, or only store one?
- Ingredient and component management, compatibility and stability tracking, and iteration across development trials
- Whether cost, composition, and compliance recalculate when an ingredient changes, or whether that is a separate step
- Multi-level formulas and sub-recipes, and whether a change at one level propagates
- Search across historical formulation data, so prior work is reused rather than rebuilt
- Whether the system can model a substitution and show its consequences before anyone runs a trial
2. How is the regulatory content maintained?
- Which markets are covered, and whether coverage matches every market you sell into
- Whether it is a maintained service or a static database, and the stated update cadence
- Who maintains it, in how many languages, and whether it is built in-house or licensed from a third party
- How a regulatory change reaches the formulas it affects, and how quickly
- Labeling and specification output, including ingredient statements and nutritional panels in local languages
3. What happens at the two handoffs?
- From R&D to manufacturing: whether an approved formula transfers automatically, with batch management, scale-up, and bill of materials synchronization
- From supplier to formula: whether raw material specifications and compliance documents are read and structured by the system, or retyped by a person
- Whether approval workflows are structured or rely on email
- Where the system exports rather than integrates, since that is where formulation data is lost and rework begins
4. What does implementation actually involve?
- The ratio of configuration to custom development in the vendor's three most recent comparable deployments
- Typical timeline for a deployment of your size and market footprint
- Who performs the work — vendor, partner, or your own team — and what that assumes about your capacity
- Service coverage in every region you operate in, and in which languages
- How far the system configures without code, and whether your team can do it
5. Will the vendor still fit in five years?
- What shipped from the last two published roadmaps, not what is on the current one
- Which AI capabilities are in production today, stated separately from what is announced
- Ownership structure, financial stability, and how acquired products were integrated
- Whether your industry is a core market for them or an adjacent one
- Position on emerging requirements such as digital product passports and lifecycle assessment
6. What do comparable customers report?
- Reference customers of similar size, portfolio complexity, and market footprint
- Whether those references use the modules you are buying, in the regions you operate in
- Retention, and the stated reasons customers have left
- What the reference would do differently if they implemented again
How you weight these depends on your timeline. A buyer replacing a failing system weights the first three hardest. A buyer making a ten-year platform decision cannot afford to discount the last three.
What Three Questions Should You Ask in the Demonstration?
|
Three scenarios separate platforms that manage formulation from platforms that only store it: a regulatory change, an approved formula moving to manufacturing, and an incoming supplier specification. Ask the vendor to perform each one live rather than describe it. |
1. A regulator restricts an ingredient today. Show me every affected product, in every market, and the labels that now need amending.
A formulation platform answers with a portfolio query. Everything else answers by describing a report you would build yourself.
2. R&D has just approved a reformulated product. Show me what manufacturing receives, and how.
This is where formulation data is most often lost and where rework originates. Watch for the word "export."
3. A supplier has sent a new raw material specification. Show me what happens to it.
The answer is either that the system reads and structures it, or that a person retypes it. Both are legitimate answers. Only one scales.Ask for these live, on the vendor's own demonstration data. A vendor that cannot perform them in a demonstration will not perform them in production.
What Are the Red Flags?
-
Regulatory content is a static database rather than a maintained service, with no stated update cadence
-
Every reference customer is in a different industry or an order of magnitude smaller than you
-
AI capability is described entirely in future tense
-
The implementation estimate assumes no integration work
-
The demonstration environment cannot be shown with your own sample formula
Nobody can tell you what happens on the day a regulation changes -
The vendor cannot say which of its last three comparable deployments needed custom development, or how much
Frequently Asked Questions
What criteria should I use to evaluate formulation software?
Six areas, ordered by how early they can end an evaluation: whether the platform models a formula rather than storing one, how its regulatory content is maintained, what happens at the R&D-to-manufacturing and supplier-to-formula handoffs, what implementation actually involves, whether the vendor will still fit in five years, and what comparable reference customers report.
How long does a formulation software evaluation take?
Timelines vary with portfolio complexity and the number of markets involved. The variable that most often extends an evaluation is scope: shortlists that mix formulation platforms with ERP, MES or LIMS take longer because the systems are not comparable.
Should I evaluate formulation software separately from PLM?
It depends on whether you need formulation science or lifecycle governance. Some platforms deliver both. Generic PLM built for discrete manufacturing does not understand formulation logic, which is why formulation management is treated as a software category in its own right.
What is the biggest hidden cost in a formulation software implementation?
Integration and configuration, not licensing. Ask every vendor for the ratio of configuration to custom development in their three most recent comparable implementations.
Is formulation management software the same as PLM for food and beverage?
Food and beverage manufacturers usually reach this category while evaluating PLM. The distinction is that generic PLM built for discrete manufacturing tracks a bill of materials, while formulation management understands what an ingredient substitution does to an allergen declaration, a nutritional panel and a cost threshold. Several vendors sell both, which is why shortlists mix the two.
Which vendors should be on a formulation software shortlist?
Vendors reach formulation from four directions: process PLM, enterprise resource planning, laboratory informatics, and scientific modeling. Decide which starting point suits your problem, rule out systems that are not formulation-aware for your sector, and compare within that group rather than across all four.
Saïda Aït
VP Global Marketing at Trace One. Saïda leads global marketing across the food and beverage, cosmetics and chemical manufacturing sectors, working directly with the industry analyst community covering product lifecycle management and formulation technologies. She works directly with the industry analyst community covering product lifecycle management and formulation technologies.
Technical review: Federico Fontanella, PMP, Head of Strategic Innovation and Product Partnerships, Trace One.