A recipe tells you what should happen. A batch record shows what actually happened. Growing food and beauty brands need both.
When orders are small, a founder may remember which supplier lot was used, why a mixing time changed, or which jar shipment caused sealing problems. That memory becomes unreliable as products, team members, and production runs multiply. Recipe and batch management software can help, but only after the business defines the records, controls, and decisions the system must support.
Food and cosmetics are subject to different regulations, and requirements vary by product, process, facility, and market. This article provides an operating framework, not legal or regulatory advice. Confirm the rules that apply to your business.
Separate the master recipe from the batch record
The master recipe or formula is the approved instruction for making a product. It should identify the version, target yield, ingredients, quantities and units, sequence, equipment, processing parameters, in-process checks, packaging, and expected output.
The batch record is the evidence for one production event. It should capture the master version used plus actual details such as:
- unique batch or lot identifier;
- production date and responsible operators;
- ingredient names, supplier lots, and quantities used;
- actual time, temperature, mixing speed, pH, or other process values;
- actual yield, waste, and number of saleable units;
- packaging component lots when relevant;
- quality results, deviations, disposition, and release decision.
Do not overwrite an old recipe when a formulation changes. Approve a new version with an effective date and preserve prior versions. Otherwise, a record saying “made according to recipe” becomes ambiguous six months later.
Design identifiers that people can use correctly
Every batch needs a unique identifier that travels with its record, samples, work-in-process, and finished goods. It may include the date or product code, but avoid a format so complicated that operators improvise.
Ingredient lot traceability should support two directions:
- Given a finished batch, identify the ingredient lots and packaging used.
- Given an ingredient lot, identify every finished batch that received it.
This matters when a supplier reports an issue, a customer complains, or a withdrawal or recall must be evaluated. The FDA’s Food Traceability Rule adds specific traceability record requirements for certain foods on the Food Traceability List; businesses should determine whether and how the rule applies to them. Even outside that rule, clear lot relationships improve investigation speed.
Test the identifier system by selecting a recent finished unit and tracing it backward. Then select one ingredient lot and trace it forward. Record gaps rather than filling them from memory.
Control recipes as governed documents
Recipe changes affect more than the ingredient list. A new preservative, flavor, pigment, container, fill weight, processing temperature, or supplier may change cost, labeling, quality checks, shelf-life assumptions, allergen considerations, or production behavior.
Use a change record that explains:
- what is changing and why;
- products and materials affected;
- tests or review completed;
- label, packaging, specification, and cost impact;
- who approved the change;
- effective date and disposition of remaining old materials.
Only approved versions should be available for routine production. Drafts should be clearly marked. Printed copies need a control method so an old sheet does not remain beside the mixer after the digital master changes.
For cosmetic businesses, the Modernization of Cosmetics Regulation Act of 2022 expanded FDA authority and introduced requirements that may include facility registration, product listing, safety substantiation, serious adverse event reporting, and records access, subject to statutory details and exemptions. Formula and batch records can support those responsibilities, but software does not itself establish compliance.
Record actual quantities and yield
A production plan may call for 100 units, but the run could yield 94 saleable units. If inventory and costing use the planned yield, finished stock will be overstated and unit cost understated.
Record actual ingredient quantities, bulk yield, filled quantity, rejected quantity, samples, spills, and remaining work-in-process. Define where normal loss occurs. Food production may lose weight during cooking or transfer. Beauty products may remain in hoses or mixing vessels. Packaging defects can reduce the final saleable count even when bulk yield is correct.
Compare actual yield with a reasonable expected range. An exception should prompt investigation, not automatic blame. Possible causes include scale calibration, incorrect unit conversion, unrecorded samples, fill-weight drift, equipment residue, or a recipe instruction that does not match real practice.
Build quality checks into the workflow
Quality data is most useful when collected at the point of work and tied to the batch. Choose checks based on product risk and specifications, not a generic template.
Possible checks include:
- ingredient identity and condition at receipt;
- time, temperature, pH, viscosity, appearance, aroma, or texture;
- net weight or fill volume;
- seal integrity and container compatibility;
- label version, lot code, allergen statement, and date marking;
- finished sample review and release status.
Define the method, frequency, instrument, acceptable range, and response to failure. “Check pH” is incomplete without the calibrated method, target, and action when the result is outside the limit.
Keep failed or pending batches separate in both physical storage and system status. A quantity should not become available for sale merely because production was recorded. Someone with defined authority should release, rework, reject, or otherwise disposition it.
Evaluate software with real exceptions
When comparing recipe and batch management software, use representative data and run one product through receipt, production, quality review, release, and traceability.
Test scenarios that expose weak designs:
- a recipe bought in kilograms but used in grams;
- an ingredient shared by several products;
- a supplier lot split across multiple batches;
- a batch with lower-than-expected yield;
- a packaging change without a formula change;
- a deviation that requires approval;
- rework using material from an earlier batch;
- an ingredient-lot search across finished goods.
Confirm who can create, approve, revise, or delete recipes and records. Look for audit history, electronic signatures where appropriate, attachments, exports, backups, and permission controls. If internet access is unreliable in the production area, understand offline behavior and recovery.
Kerno brings formulas, inventory, production records, costing, and QA into a connected product workflow. That connection can reduce duplicate entry for a growing small-batch brand, but the business must still define its specifications, train operators, verify data, and understand applicable food or cosmetic requirements.
Start with one controlled product family
Implementation should be narrow enough to verify. Clean item names and units, assign ingredient and finished-batch identifiers, approve one recipe, and enter opening inventory from a physical count. Run several batches and reconcile expected versus actual consumption and output.
Then perform a mock trace. Choose a finished batch, locate all inputs and quality results, and identify where every saleable unit went if distribution data is maintained. Choose an ingredient lot and locate all affected batches. Time the exercise and fix the gaps.
Good batch management makes four facts clear: what should have been made, what was actually made, what went into it, and whether it was approved for sale. Establish those facts first. Software should make the control easier to execute and inspect—not replace the thinking behind it.
