Signs Your Supplement Manufacturer Is Failing, and What Switching Really Involves

Reviewed by Apollo’s quality team · Last reviewed July 2026

A run date slips. Then the next one. A status request goes a day without an answer, then two. A batch comes back a little off (the flavor isn’t quite right, or the fill runs light) and the reply is “we’ll look into it,” which is not the same as an answer. Somewhere in there you stop assuming it’s a rough patch and start wondering whether the shop itself is the problem.

Here’s the distinction that decides it, and it’s the one most “time to switch” lists skip. Almost every manufacturer has a bad week: a missed date, one off-spec lot, a slow reply. Those are hiccups, and hiccups get fixed. What you’re actually diagnosing is whether the failure is structural: baked into how the shop is built, which means it repeats, and no phone call resolves it. This is that diagnostic, written from the receiving end.

Educational overview: not legal, regulatory, or medical advice. Requirements change and vary by jurisdiction and sales channel. Last reviewed July 2026.

Short answer. Almost every manufacturer has a bad week (a slipped date, one off-spec lot, a slow reply) and hiccups get fixed. A failure is structural when it repeats and no one can explain it against a written standard: run dates that keep slipping with no root cause, status that goes silent under pressure, quality that drifts with no spec to measure against, or price and minimums that climb with no cost driver. Structural failures don’t improve with another phone call, because they’re how the shop is built.

Best for: Brands feeling delays, batch drift, opacity, or reorder friction and deciding whether to stay or move.

Key decision: Whether the failure is a fixable hiccup or a structural one, and (if it’s structural) how to exit without a stockout.

Apollo path: The review-first transfer playbook is the exit that protects supply; a transfer review reads your product and returns a benchmark-sample plan.

Apollo Future Labs takes in products from manufacturers that stopped keeping up (we run the benchmark, the receiving review, and the pilot on the incoming side) from an FDA-registered facility in Livermore, California, with cGMP-compliant operations on the lines. So the failure signatures below aren’t a listicle. They’re what we watch walk in the door.

Four surfaces, one question: hiccup or structural?

A manufacturing relationship fails on one of four surfaces, and only four: the schedule (do runs happen when they’re supposed to?), communication (do you know what’s happening without chasing?), quality (is the product right, and can they prove it?), and the commercials (are price and minimums stable and explained?). Every symptom you can name lands on one of them.

The diagnostic isn’t the symptom; everyone has symptoms. It’s whether the symptom has a root cause the shop can name, a written standard behind the answer, and a frequency of once. Run any failure through that and it sorts itself into fixable or structural.

Each section below reads one surface: the fixable version, the structural version, and the question that separates them.

When run dates stop being reliable

A single missed run is a conversation. Materials came in late, a line went down, a tech was out; it happens, and a shop running a real schedule tells you before the date, names the cause, and gives you a recovered date that holds. That’s a hiccup. Annoying, not terminal.

The structural version is a pattern. Dates slip, then slip again. The new date is offered confidently and then slips too. And when you ask why, you get weather (“we got behind,” “it’s been busy”), never a root cause tied to something that changed. That pattern usually means one of two things, and neither improves with a call: the shop has overcommitted its lines and you’re the account that absorbs the overflow, or it has hit a capacity ceiling and can no longer fit your run without displacing someone else’s.

The tell is specificity. A shop in control can tell you exactly why a date moved and what’s different about the next one. A failing shop can only tell you it’s sorry, and sorry doesn’t ship product. Track three run dates: if the reason is always vague and the slip is always a surprise you discovered rather than one they disclosed, the schedule isn’t having a bad month. It’s structurally underwater.

When status goes quiet under pressure

Communication is the earliest warning surface, because it fails before the others become visible. The pattern worth watching: a healthy shop’s communication is roughly symmetrical. You ask, they answer; and when something goes wrong, they call you. A failing shop’s communication is one-directional. You chase; they respond, eventually. And when something goes wrong, the line goes quiet; you find out a run slipped because the tracking number never came, not because anyone told you.

That asymmetry is diagnostic on its own. Silence under pressure is almost never a personality trait. It’s usually the sound of a shop managing a problem it doesn’t want to disclose: a batch that didn’t pass, a material that didn’t arrive, a schedule it has already lost. The opacity isn’t the failure; it’s the cover for the failure on one of the other three surfaces.

So treat a communication pattern as a lead indicator, not the whole diagnosis. A shop that goes dark exactly when you most need an answer is telling you something is wrong upstream (on the schedule, in a batch, or in its own capacity) that it has decided not to say out loud. The question that cuts through it: when there’s bad news, who says it first? If the honest answer is “always me, after I go looking,” you’re not being kept informed. You’re being managed.

When quality drifts, and nothing measures it against a spec

This is the deepest surface, and the one a brand can’t fix from the outside. Every product on your shelf has a right answer: a written specification that says what a correct batch contains and how it behaves (identity, purity, strength, composition, and the physical attributes a customer notices). A working quality system tests each batch against that spec and keeps the record. When quality “drifts,” what’s really happening is that batches stop landing on the spec, and a failing shop often can’t tell you by how much, because there was never a firm spec to land on.

That is not a rare failure. FDA inspection data has repeatedly shown that a specifications failure ranks among the most-cited problems in supplement manufacturing: the shop either never established written product specifications for identity, purity, strength, and composition, or never verified that a finished batch met them. Trade summaries of FDA enforcement data, such as Eurofins’, have tracked the pattern for years, and the rule itself (FDA’s 21 CFR 111.70 and 111.75) requires both the specifications and the testing to confirm them. A shop without them isn’t unlucky. It’s built on sand.

This is what the drift looks like from your side of the counter, and the spec that should have caught it:

Two specific tells separate a real quality system from a badge on a website. The first is single-ingredient testing: a shop that assays one active in a multi-ingredient product and calls the batch verified isn’t meeting the standard; 21 CFR 111 expects the finished batch to meet its specifications, not one component of it. It’s a documented citation pattern, not a gray area. The second is records that don’t reconcile: the master manufacturing record says one thing, the batch production record says another, and the Certificate of Analysis is from a lab nobody will name. When the paper doesn’t line up, the product isn’t controlled; it’s improvised, and the next batch is a coin flip.

So make the request that ends the diagnosis: ask for the specification a given batch was tested against, and the batch record and third-party CoA that prove it. Drift you can measure, sitting on top of a spec that doesn’t exist, is structural. No phone call adds a quality system that was never built. This is the failure most often riding along on the transfers we receive, precisely because it’s the one a brand can’t repair from the outside; it’s the difference between the specification a real quality system keeps and a shop that’s been guessing.

When price and minimums move without a reason

The commercial surface fails two ways, and both hide behind the same tactic: a number you can’t decompose. A legitimate price change traces to a named input: a raw-material cost you can verify moved, a new test you actually asked for, a freight lane that repriced. A legitimate minimum traces to a component supplier’s lot size: custom labels, bottles, closures, and droppers commonly run in the low thousands, set by the printer or molder, not by your manufacturer. In both cases the number is a line item you can see and challenge.

The structural failure is the number that floats free of any line item. Price creep shows up as quotes that climb quarter over quarter with no input you can point to: a shop repricing you because switching is hard and it’s betting you won’t. Reorder MOQ walls show up as a minimum that suddenly jumps at reorder, or a blended “MOQ” that mixes production and components so you can’t tell which is driving it. Sometimes that wall is a shop that has hit its own scaling ceiling and is quietly managing itself out of your small-to-mid runs; sometimes it’s a shop that never separated the two numbers to begin with.

The move is the same either way: make them decompose the number. A finished-unit run and its component minimums are two different figures with two different owners, and a shop running an honest quote will show them on separate lines; that’s how a legitimate minimum is actually built. If the price can’t be traced to an input, or the minimum can’t be split into production versus components, you’re not looking at a cost. You’re looking at leverage, and leverage that grows every quarter is a structural signal, not a negotiation.

What each failure actually costs you

The reason to diagnose this precisely is that the cost of a structural failure doesn’t sit still. It compounds, and it compounds downstream where it’s hardest to recover: in your channel, your ranking, and your margin.

Two of these deserve emphasis because brands consistently underprice them. A stockout on a ranked listing isn’t a pause in revenue; it’s a reset, because the rank and the review velocity you spent months building don’t wait for you, and reclaiming them costs more than the lost sales did. And discounting off-spec stock to clear it doesn’t just cost the margin on that batch; it trains your customers to wait for the discount. The failure that looks like an operations problem lands as a marketing and retention problem, one channel over.

Fixable or terminal? The decision gate

You don’t need a scorecard. Three questions decide it, and you run the failure, whichever surface it showed up on, through all three:

1. Can they name the root cause? A shop with a working system can tell you why it happened and what’s now different. “We’ll look into it,” repeated, is not a root cause.
2. Is there a written standard behind the answer? A specification, a batch record, a documented input change. If the answer floats free of any document, there’s nothing underneath it to hold.
3. Did it happen once, or is it the pattern? One event with a fix is a hiccup. The same failure a third time is how the shop is built.

Yes, yes, once means stay: hold them to the corrective action and watch whether it holds. No, no, pattern means it’s structural, and structural doesn’t get better with another call. That combination is your exit signal. It’s also the honest floor for a hard conversation: a shop that can answer those three questions plainly is one you can work with; a shop that can’t is one you’ve already outgrown, whether or not you’ve admitted it yet.

What switching really involves: the honest cost

Switching is not a swap, and any page that makes it sound like one is selling you something. It’s a project with a real cost, and you take it on when the cost of staying (the compounding churn, rank loss, and discounted inventory above) exceeds it. Knowing the true cost is what lets you make that call instead of flinching from it.

Here’s what it actually involves. You run two relationships for a while, on purpose. You keep your incumbent shipping while a new shop earns your supply, because the danger in switching isn’t picking the wrong shop; it’s going dark in the middle and stocking out on a product that was selling. It takes weeks, not days. Getting a new shop from a benchmark sample to a passed pilot run typically runs about six to eight weeks; a formula, packaging, or process change can pull in stability work that extends the full cutover to weeks-to-months. If that change re-opens your shelf-life dating, you re-establish the data before the new product carries your expiration date, and you hold safety stock through that retest rather than pausing supply.

Two costs surprise people, so name them before you start. You may not own your formula. On a private-label or stock base, the recipe can belong to the incumbent, not the brand on the label; you keep your brand and usually your artwork, but the formula may need rebuilding by benchmark-matching or custom work. Confirm ownership before you plan the move, not during it. We support the operational side of a transfer; your counsel decides the contract-exit and ownership questions. And the overlap itself costs cash: extra finished-goods cover held, an incumbent kept warm. That’s not waste; it’s the premium you pay to keep the switch off your live supply until the new line is proven (what makes a transfer boring instead of a bet).

The exit that protects your supply

The exit that works is the opposite of a hard cutover: you prove the new shop before you let it near your live supply. In one line, the shape is a benchmark sample matched to a unit you sell today, a spec-and-component review that maps where quality would drift, a pilot run on the real production line (often in the hundreds of units when compatible materials are on hand), and then a cutover you overlap with existing stock, so the incumbent keeps shipping until the new line has shipped a clean run at volume. Nothing about your supply moves until the new shop has earned the move.

That’s a full playbook, and rebuilding it here would be noise: the review-first transfer playbook walks each gate, the stability trigger, and the cutover checklist. This piece is the diagnosis; that one is the execution: the review-first contract manufacturing we run as a specialty, and the path we built for brands mid-switch.

One thing worth knowing before you send anything: a fast no is a service. If you bring us a product that isn’t a fit for our lines, you’ll hear that in days, not strung along for weeks, because a fast no sends you looking elsewhere while your incumbent is still shipping. A transfer review typically comes back in one to two business days, with a real read on your product either way.