Viewpoint

Past Its Used-By Date

Software does not spoil. It does not smell, or separate, or change colour. It simply goes on looking exactly as it did on the day you bought it, long after it stopped being right.

Everyone in this industry understands shelf life intuitively, and understands that it is not really about a date.

A product does not become unsafe at one minute past midnight. It degrades continuously, gradually, and invisibly, from the moment it is made. The date on the pack is not the moment of failure. It is simply the point at which the business agrees to stop pretending otherwise. The whole discipline exists because degradation is silent, and because silence is not the same as safety.

Apply that lens to the tools your team relies on.

Compliance platforms have a shelf life too. They are built to answer the regulatory conditions of a particular moment, and those conditions decay from underneath them. But software gives you no sensory warning that this has happened. There is no smell, no separation, no change in colour. The interface looks the same. The digest arrives on schedule. The invoice is unchanged. The tool bought in 2016 to solve the problem of 2016 continues, in 2026, to present itself with total confidence as the answer to a question nobody is asking any more.

There is no used-by date printed on a regulatory intelligence platform. Which means it is entirely possible to be running one that expired years ago, and to have no way of knowing except by what keeps going wrong.

This piece is about how to read those signs, and about the specific, expensive, and well-documented ways that an out-of-date tool fails a food business. Not in theory. On the shelf.

1 It was fresh once

Fairness first, because it matters to the argument.

When these platforms were bought, they solved a genuine and urgent problem. Regulatory information was scattered, multilingual, buried, and slow to surface. A rule could change in a market you sold into and nobody in your business would hear about it for months. That was a real risk, and the remedy was the right one: build a large, comprehensive, central source that made it impossible for a change to happen without you seeing it.

Visibility was the constraint, and breadth was the cure. Buying it was not a mistake. It was competent.

But that constraint has been solved, comprehensively, and for some years now. And a cure that outlives the disease does not become neutral. It becomes an ingredient with its own effects.

Ask any regulatory affairs professional in food and beverage what actually keeps them up at night, and almost nobody says "I am worried a regulation might change without us hearing about it." They say the opposite. They say that forty things changed, thirty-eight are irrelevant, and there is no fast, defensible way to establish which two are not, or what those two mean for a product that is already in production.

The problem moved from finding out to working out. That is not a harder version of the same problem. It is a different problem, and breadth is no answer to it at all.

2 Nothing that went wrong was invisible

Here is the part that should give any Head of Regulatory pause, because it is not a projection or a scenario. It is what the recall data already says.

There were 567 food and beverage recalls in the United States in 2025, up from 513 the year before. Undeclared allergens accounted for 261 of them. Which means, as the industry’s own annual review puts it plainly, that nearly half of all recalls trace back to labelling failures rather than contamination. Not to a pathogen. Not to foreign matter. To a label that did not say what it needed to say.

And the average recall is getting larger, not smaller. Analysis of the most recent quarters found the average size of a US food recall nearly doubled, reaching 57.4 million units, even as the total number of recalls fell. Fewer events, each one bigger.

Now consider how an allergen recall actually happens, because it is almost never dramatic, and it is almost never a failure of knowledge.

Read that again, because it is the whole argument in miniature. That is not a discovery failure. No horizon scanning platform on earth would have prevented it, because there was nothing to discover. The regulation was known. The change was recorded. The failure occurred in the gap between knowing a rule and establishing, quickly and accountably, what it meant for a specific product on a specific line on a specific day.

That gap is where food businesses are losing money. And it is precisely the gap that a tool optimised for breadth does not address, cannot address, and was never built to address.

See where that gap sits in your own portfolio → Get in touch

Look down the middle column. In every case, the business already knew. The information was not missing. The state bills were public, the effective dates were published years in advance, the allergen rules are decades old, the claims environment has been contested for a generation.

The next compliance failure at your business will almost certainly not be something nobody saw coming. It will be something everybody saw coming, that nobody could turn into an answer in time.

This is why the question is not really whether the platform is good. It may well be excellent at what it does. The question is whether what it does is the thing that is currently costing you money.

3 What the exposure actually looks like

It is worth being concrete about the downside, because it is not abstract and it is not slow.

The recall

A recall is never merely a logistics exercise. It requires formal notification, product retrieval, root cause investigation, and corrective action reporting. Direct costs for a large-volume product run into millions before anybody has calculated the value of the lost product itself. One USDA analysis of a single ground beef recall put the indirect losses, from reduced consumer purchasing alone, at over ninety-seven million dollars across the industry.

The retailer

This is the consequence that regulatory teams tend to underweight, and commercial teams never forget. Labelling and packaging non-compliance produces chargebacks, rejected deliveries, and missed launch windows. Retailer scorecards record it. Buyers remember it. A brand that repeatedly creates receiving friction finds that future listings become harder to win, even when the product itself performs. Delisting is the ultimate expression of that, and it is used routinely as leverage.

The window

And then there is the quietest cost of all, which never appears in any incident report. The product that could have launched in a market but did not, because nobody could confirm the claim in time and the season passed. Nothing failed. Nothing was recalled. The revenue simply never arrived, and no one wrote it down.

Set that exposure against what is actually being spent. Regulatory intelligence platforms in this category run to roughly fifteen to twenty-five thousand pounds per user per year at the mid-market, with enterprise agreements comfortably into six figures. That is a substantial sum, and the question worth asking is not whether it is a lot of money. It is: what, precisely, is it buying protection against? If the answer is "the risk of not hearing that something changed," then the spend is well-defended against a risk that has largely been solved, and structurally undefended against the one that is currently generating the recalls.

Not sure your spend is buying the right protection? Get in touch.

4 Can you afford to carry it another year?

This is where the shelf-life analogy earns its keep, because it tells you something uncomfortable and true about time.

A product past its date does not hold steady. It continues to degrade, and the longer it is held, the worse the position becomes. Nobody in food believes that keeping something on the shelf makes it fresher.

The same is true here, and in two directions at once.

The regulatory environment continues to fragment. The state additive patchwork is growing, not consolidating. Texas now requires on-pack warnings for products containing any of forty-four specified dyes or additives. Louisiana requires QR codes linking to ingredient disclosures. Fifteen states introduced ultra-processed food bills in a single year, with more expected. Each new rule widens the gap between knowing what changed and knowing what it means for your portfolio, which is precisely the gap your current tool does not cover.

And at the same time, the tool embeds further. Another year of workflows built around it. Another year of reports that assume it. Another year in which the manual filtering process your team performs on top of it hardens into something that looks like the job.

Every year you carry it, the exposure grows and the cost of putting it down grows with it. There is no year in which this becomes easier. This one is the cheapest it will ever be.

So the question, put plainly, is not whether you can afford to change. It is whether you can afford to spend another twelve months structurally undefended against the failure mode that is currently producing half the recalls in your industry, while paying a six-figure sum for protection against a failure mode that has largely ceased to exist.

Framed that way, "we will look at it next year" is not the cautious option. It is the expensive one, and it is expensive in a currency that shows up on the shelf rather than in the software budget.

"Nobody in this industry would keep an ingredient past its date because throwing it out felt like admitting they had bought it wrongly. They would throw it out, because they understand what happens if they do not. We are strangely willing to apply a different standard to the tools we use to keep the food safe."

— Nicola Colombo, Co-founder, Prodeen

Nicola spent two decades in this category, including the years when a bigger database genuinely was the right answer. He is not describing somebody else’s mistake. He is describing an industry that stopped checking.

"Ask what your platform actually prevented last year. Not what it sent you. What it stopped. If the room goes quiet, you have your answer, and it did not require an audit to find it."

— Nicola Colombo, Co-founder, Prodeen

5 Checking the date

None of this requires anybody to concede that the original purchase was wrong. It was not. It requires only the ordinary discipline that this industry already applies to everything else it holds: check the date, and act on what you find.

Four questions will establish where you stand. They can be asked in an afternoon.

1. What did the platform prevent in the last twelve months? Not what it delivered. What it stopped. If nobody can name an instance, that is not a criticism of the team. It is a measurement of the tool.
2. When something went wrong, or nearly did, was it because nobody knew, or because nobody could establish what it meant in time? Almost everyone answers the second. That answer tells you which problem you actually have.
3. What proportion of the coverage you pay for intersects with the products you sell and the markets you sell them in? Ask the incumbent directly. The manner of the answer is as informative as the answer.
4. When does the notice window close? Not the renewal date, which is later and mostly decorative, but the notice date that precedes it, usually by thirty to ninety days. Around three quarters of software vendors rely on automatic renewal, and price increases of ten to twenty per cent are routine. If nobody can tell you that date within the hour, the decision is currently being made by a calendar.

If the answers are reassuring, renew with confidence. That is a legitimate outcome and this piece does not argue otherwise.

If they are not, the useful next step is evidence rather than instinct. Prodeen runs a Coverage Audit: we map your real product portfolio, market by market and category by category, against the coverage you are currently funding, and against the specific interpretive gaps where the recalls are actually occurring. It runs while your existing contract is live, it carries no obligation, and it is built to be shown to a CFO. If it shows your current arrangement is well matched to your portfolio, we will tell you so, and you should renew. We would rather report that than take on a customer we are not suited to serve.

What it produces, either way, is a date. Something the industry has always understood how to act on.

Every business in food knows what to do with something that is past its date. The only unusual thing here is that nobody printed one on it.

Sources and further reading

This piece draws on published US recall data for 2024 and 2025, on current food and beverage regulatory analysis, and on procurement and organisational research. Recall figures are US-based and are cited as indicative of a pattern rather than as a global measure.

1. The National Provisioner (2026). "2025 recalls: the year in review." Records 567 US food and beverage recalls in 2025, up from 513 in 2024, with undeclared allergens accounting for 261 events. Notes that nearly half of all recalls in both years trace back to labelling failures rather than contamination.
2. Esko (2026). Analysis of 251 FDA food recalls from 2025. Describes the typical anatomy of an allergen recall: a formulation changes, the specification is updated, but the label artwork was created months earlier and is managed by a different team.
3. ProFood World (2026). "Food Safety Prevention: Moving Beyond Recall Response." Reports approximately 320 FDA and USDA food recall and public health alert announcements in 2025, with undeclared allergens the leading cause at roughly 39%.
4. Sedgwick (2026), reported via MySA. Found the average size of US food recalls nearly doubled between Q4 2025 and Q1 2026, reaching 57.4 million units, with undeclared allergens and labelling errors among the most common causes.
5. USDA Economic Research Service. Trends in Food Recalls, 2004 to 2013. Documents that undeclared allergen recalls nearly doubled across the period.
6. Davis Wright Tremaine (2026). "2026 Food Regulatory Update: Ingredients in the Crosshairs." Records that fifteen US states introduced ultra-processed food bills during 2025.
7. Skadden (2025). "Dealmaking in the Food Industry: Navigating State and Federal Food Regulations." Describes the rolling, inconsistent set of state compliance dates now facing suppliers, including Texas on-pack warning labels for 44 specified additives and Louisiana QR code disclosure requirements.
8. Gibson Dunn (2026). "The Year Ahead for Food and Beverage." Notes expanding litigation around ultra-processed foods, functional claims, and packaging.
9. MacMillan SCG (2026). "Why Retail Compliance Mistakes Are Costing FMCG Brands More." Describes chargebacks, rejected deliveries and missed launch windows arising from labelling and packaging non-compliance.
10. Staw, B. M. (1976). "Knee-deep in the Big Muddy." Organizational Behavior and Human Performance, 16(1). On escalating commitment to a chosen course of action.
11. Gartner research on SaaS retention, reporting that roughly 75% of vendors rely on auto-renewal, with renewal price increases of 10 to 20% common.
12. Changeflow (2026). Regulatory Intelligence Software Pricing. Indicates roughly $15,000 to $25,000 per user per year at the mid-market, with enterprise agreements crossing six figures.