Reformulation: what the recipe change sets off

Changing the formula is the smallest task on the list. Bench work on a new recipe can be done in weeks. What follows it, the specification, the validation, the panel, the pack and the person who has been buying the old version for two years, is where reformulations quietly overrun.

Brands reformulate for cost, for supply, for a texture complaint, for a claim, or because a labelling regime has made an ingredient expensive to declare. India’s food regulator has proposed a front-of-pack warning system for packaged foods, which if it proceeds will send some brands to the bench rather than to the printer. The sequence below applies whichever reason brought you here.

The recipe is the cheapest part

A formula that works at five kilograms on a bench does not automatically work in a production vessel. Scale changes shear, heat transfer, mixing order, hold time and how the product behaves in the filler. The first honest test is a line trial at the plant, run on the actual equipment, with the actual pack.

Budget for at least two trials. The first tells you what you got wrong and the second tells you whether you fixed it. Planning for one trial is planning to launch on the first attempt.

Reissue the specification, not just the formula

The failure mode is boring and expensive. The plant fills to a new formula while quality releases against the old specification sheet, and nobody notices until a result sits outside criteria that were never meant to apply.

Reissue the whole document set together, with a version number and an effective date every party works to.

• Finished product specification and its release criteria.
• Bill of materials and the ingredient declaration in the order it will appear on pack.
• Process document with any changed parameters.
• Quality control methods for anything the change touches.
• Incoming material specifications for anything new.

Every document should name the version it replaces. If your co-packer is working from a PDF someone emailed last year, you do not have a specification, you have a rumour. The discipline that carries an audit is what carries a reformulation.

Validation is the lead time

Shelf life is not inherited. The previous study was run on a different product, and a change to a fat, an acid, a preservative system, a water activity or a pack material can change how the product ages in ways the old data cannot predict.

Treat it as a process rather than a date. Define what you are measuring before you start, across sensory, microbiological, chemical and physical attributes. Agree acceptance criteria in advance, because criteria written after you have seen results are not criteria. Run against a control, which is the current product. Test in the pack you actually sell, because the pack is part of the system. And run real-time work alongside any accelerated work, since accelerated data is an indicator and not a substitute.

How long that takes depends on the claim you want to make and the risk profile of the product, and it is not something you can compress by wanting to. What your category requires, and what evidence has to sit behind any shelf-life statement, is a question to settle with your own regulatory advisor before the first commercial batch.

If the panel changes, the pack changes

The moment declared values or the ingredient list move, the artwork moves with them, and you have acquired a second project. That project has its own inventory arithmetic: committed film, weeks of printed cover and the run-out versus scrap decision, all of which is its own exercise. What the panel has to carry is covered in nutrition and allergen labelling, and any claim the reformulation makes newly possible, or newly impossible, belongs in the same review as the rest of your pack claims.

Sequencing matters here. If the panel change is confirmed, brief the artwork and the print buy in parallel with validation rather than after it, so the lead times overlap instead of stacking.

A repeat buyer is judging you against memory

The repeat buyer did not ask for this and does not care why you did it. They will taste the new version against a memory of the old one, and if the gap is large enough they will simply stop buying and never tell you.

Find out before the market does. Run blind paired comparison against the current product with people who actually buy the SKU, not colleagues. Separate two questions that get collapsed: is the difference detected, and is it disliked. A detected but accepted difference is manageable. A detected and disliked difference in a product bought on habit is a revenue event. Getting the sample and the questions right matters more than sample size, which is why study design decides whether the answer is worth anything.

After launch, watch review text and support tickets for taste, texture and smell words against a pre-change baseline you captured deliberately. Without it you cannot tell a spike from normal noise.

New ingredient, new supplier, new qualification

Anything new in the bill of materials brings a vendor who has never shipped to you, documentation you have not reviewed, incoming checks you have not defined, and the question of whether you need a second approved source for anything the formula cannot run without. That is a full workstream in itself and worth running to the same standard as any other supplier qualification. The change will also move your cost sheet, usually in more lines than the one ingredient you swapped.

Is it still the same SKU?

Commercially, you are choosing whether to carry the change on the existing code or open a new one. Keeping the code preserves review history, ratings, rank and buyer records. Opening a new code is clean but restarts all of that from zero.

Most brands keep the code, and that is usually right. The cost is that the code no longer tells you which version a unit is, so you need another way to separate them in your own data while both ship, which is the transition window and its own piece of work. Whether the identifier itself has to change is a separate technical question answered under identifier hygiene.

One line to keep in front of the whole project. The regulatory consequences of a formulation change, including anything that has to be notified, declared or evidenced, are for your own regulatory advisor to determine for your product and your category. Do not run a reformulation on what a competitor appears to have done.