01 Start with the CCPs your product actually requires

Before you build anything, you need to know which CCPs apply to your product. That is what tells you what has to be recorded and when. You cannot design a log for checks you have not identified yet.

This varies by what you make. For a lacto-fermented sauce, the controls center on pH: the pH of the sanitizing solution for your bottles, and the finished equilibrium pH of the sauce itself. For a koji sauce that gets a thermal step, you are watching pasteurization time and temperature and running a hot fill. Different process, different critical points, different log.

Where those requirements come from depends on how your product is classified, and that classification is not a guess you should make alone. Shelf-stable, pH-controlled products often fall under the acidified foods rule, 21 CFR Part 114, which requires you to control finished equilibrium pH at or below 4.6 and to record it (§114.80). But whether a given product is legally an acidified food, an acid food, or a formulated acid food is a determination a process authority makes for your specific formulation. Get that letter. It defines your CCPs for you, in writing, which is the whole foundation of the system you are about to build.

If your product carries a thermal preventive control, like the pasteurization step on a koji sauce, you are also in the world of FSMA preventive controls under 21 CFR Part 117.

The point of step one is simple: get the list of required checks nailed down and documented before you decide how to record them.

02 Decide where the CCP record lives, and build it for the floor

Once you know what you are recording, you need somewhere to record it. Realistically there are three options: a paper form, a spreadsheet, or purpose-built software.

A paper form works until a binder gets wet, a sheet goes missing, or the writing is illegible six months later when somebody needs it. A spreadsheet works until the moment you actually need to update it during production, which is the moment it falls apart. We lived the spreadsheet version. The data was never wrong because the readings were bad. It was wrong because nobody could keep it current while also making product.

Whichever you choose, design it around one question: can the person taking the reading enter it right there, without stopping what they are doing? If the answer is no, you have already lost. Which brings us to the part that actually matters.

03 Log at the equipment, in the moment

This is the whole thing. Everything else is setup.

Retroactive logging is how records go missing and how mistakes get baked in. A check logged from memory at the end of a ten-hour shift is not a real record, it is a reconstruction. The fix is not to try harder to remember. The fix is removing every step between taking the reading and recording it.

Every step added to a process, no matter how well intentioned, is a little bit of friction. The more friction in a process, the less likely it is to be followed.

Jeremiah Utecht, co-founder, FourFoxes

That is the entire reason systems get abandoned. It is not that your team does not care. It is that texting a reading, then transcribing it later, then double-checking it, is three steps where there should be one. Every one of those steps is a place for the record to die.

So the test for any CCP system is friction. With CCP and calibration logging in FourFoxes, the way this works on our floor is: you open the batch record at the start of work, and logging a CCP is entering the value and tapping save. That is it. It is less work than sending a text message. When recording the reading is easier than not recording it, your records stay complete on their own, because the path of least resistance is the right one.

That is the bar. Whatever you build, make the in-the-moment entry the easiest possible action.

04 Track your calibration, including the why

The check most small producers underbuild is calibration. Your pH reader and your thermometers are only as trustworthy as their last calibration, and you need to track not just that you calibrated, but when and why.

Most of the time this is routine: you calibrate on a schedule, you log it, you move on. The why matters when something out of the ordinary happens. If you have to replace a probe, or a reading comes back off and you recalibrate mid-run, that is the record an inspector will actually want to see, because it tells the story of whether your monitoring was reliable when it counted.

Under the preventive controls rule, calibration of your process monitoring instruments is a required verification activity, and you are expected to have written procedures covering the method and frequency (21 CFR §117.165(a)(1) and (b)(1)). The instruments themselves have to be accurate, precise, and maintained (§117.40(f)). Build calibration logging into the same system as your CCP checks, not off in a separate notebook that gets forgotten.

What CCP tracking is really for

It is easy to treat all of this as paperwork. It is not. Every health inspection, you have to show that your CCPs were validated and hand over your calibration records. Those are not casual checks, and as you grow, the FDA can examine the same records. They have real consequences.

One more practical detail that trips people up: how long you keep the records depends on which rule governs them. Acidified foods records under Part 114 are retained for three years from the date of manufacture (§114.100(e)). Records under the preventive controls rule are retained for two years after they were prepared (§117.315(a)(1)). If both rules touch your operation, keep to the longer requirement and do not guess.

Set it up in this order. Identify the checks, decide where they live, kill the friction so they get logged in the moment, and track your calibration like it matters. Do that, and the end-of-shift scramble disappears, because there is nothing left to reconstruct.