A short, risk-weighted count each week, taken on the barcode scanners your staff already carry — so accuracy is something you hold all year rather than something you buy once.
Annual stock-takes are expensive, disruptive, and out of date the day after they finish. Manual counts drift. Shrinkage hides in the gap. There's a better rhythm.
You pay an external team, close down to count everything at once, and the figure is stale within days. The rest of the year, you're guessing.
A small, risk-weighted count each week keeps accuracy live all year — staff do it in the gaps of a normal shift, no shutdown required.
Paper sheets, discretion over what gets checked, and errors noted but rarely resolved. Discrepancies vanish into the next count.
The app dictates what to count by risk, forces every discrepancy to be resolved, and seals each session — so the numbers can be trusted.
Estiba picks each week's count by risk — weighting high-value and fast-moving stock automatically. No one has to decide what matters.
Operators count the week's items with any barcode scanner. Anything that can't be resolved on the spot is held safely, not lost.
Every discrepancy is resolved before a session closes, then signed and sealed — giving you an accuracy figure and a shrinkage cost you can act on.
Estiba is built for businesses that operate their own shops, hold stock worth enough that a discrepancy matters, and move it between locations because customers ask them to.
In practice that has meant premium and luxury retailers running a handful of shops, and businesses that have grown past the point where everyone is close enough to ask. Practice that lives in people works perfectly until the geography changes.
Where Estiba is today. The first pilot is in preparation, and I am deliberately starting with small estates, where the whole thing can be run properly and measured honestly. If you run a larger estate and this is still interesting, say so — I would rather tell you plainly what I am ready for than find out together.
Inventory tools are only worth anything if the figures are honest. Estiba is built so the number on the screen is one you'd defend to your CFO — measured, not guessed.
The count concentrates where the money and the movement are, so the same effort tells you more.
A session can't close with units unaccounted for. Every gap is found or formally recorded.
Shrinkage at cost and lost sales at retail, kept separate — no inflated single number.
Estiba runs on the barcodes already on your stock and the phones or scanners your team already carries. Roll it out across every store without buying a single new piece of equipment.
Because you still count. RFID does not remove the count; it changes what a count costs. A process that took a team a weekend takes a couple of hours, which is why estates that install it count weekly rather than annually. The reading gets cheap. The discipline still has to exist.
And the scarce resource moves. When counting was expensive, the hard question was what do we count. When counting is cheap, you are handed a variance list every week with more lines on it than anyone has time to investigate — so the hard question becomes which of these is worth someone walking over to. RFID makes that problem bigger, not smaller.
Nor does a reader resolve anything. It produces a difference; it does not hold a session open until a person explains it, record what they saw when they got there, or accumulate those explanations so that a cause surfaces instead of a correction repeating. It cannot see what was never tagged, since tagging is only economic above a certain value. And it reads presence rather than condition — a tag encoded wrongly at source is propagated with total confidence, where somebody looking at the garment would have caught it.
The useful way round is this. Estiba is the method; RFID is a very good way of feeding it. Cheap reads mean frequent reads, frequent reads mean a longer history, and a longer history is precisely what makes variance explainable rather than merely visible.
Which means the businesses that have invested most in RFID have the most to gain from a process that uses it. A reader justified on accuracy that produces a number nobody acts on has bought a measurement, not an improvement.
Estiba is built by Alejandro Gil Galue.
I priced and monitored risk for a living — corporate and investment banking, and then investments of my own — and later spent time in retail stockrooms. Not as a career, but as someone who could not stop looking at them as a system.
What struck me was not disorder. The people were quick, the shelves were organised, the process was followed — and all of that was being read as evidence that the operation was sound. Appearance had become the standard by default, because nothing was in place to assess anything else. You can look tidier. You cannot be more correct.
Anyone who has priced risk recognises the shape of that immediately: an exposure that is real, continuous and unmeasured — and therefore unmanaged, because nothing unmeasured is ever managed.
Estiba is built to be used by whoever is actually standing in the stockroom, which is a design constraint rather than a sentiment. A count is only worth what the person taking it was in a position to notice.
It is a small company and I am the whole of it, which means anyone who works with Estiba works with the person who built it. I regard that as the point rather than a limitation.
What counts as success is agreed before anything starts.
A baseline is measured at go-live. We set a threshold together — the improvement that would make this worth doing — and we fix the method for measuring it in advance, on a sample drawn independently of the items the system chooses to chase. A system that grades itself on the work it selected is not being graded.
Then a quarter. At the end you get the figures, measured that way, whether or not they flatter the product.
If the threshold is missed, the pilot ends there and nothing further is owed. If it is met, we will have agreed beforehand what happens next — which is the whole point of agreeing it beforehand, and it binds me as much as it binds you.
Request a pilot →