By making it too bulky to conceal, while keeping it visible - and by locking so it cannot be opened away from the till.

The product itself is small enough to pocket, which is why it is targeted. The case removes that property.

Transparency is essential: the customer must be able to read the packaging, or the protection costs a sale.

The lock is opened with a key at the point of sale, and the product is handed over without the case.

Most cases carry an EAS element as well, so an item taken in its case alarms at the exit.

The visible presence of safers on a line is itself a deterrent, which is part of the value.

Closely - a product that can be shaken about inside the case can often be worked out through the opening, and a loose case gives false confidence.

Safers come in a range of sizes and stores generally need several to cover their protected lines.

Measure the actual products, including any variation between pack sizes in the same range.

Packaging changes are the usual cause of a fit going wrong: a supplier redesigns a carton and the existing cases no longer suit it.

Check the closure and hinge design as well as the internal dimensions - some patterns leave a gap at the hinge.

Audit fit periodically rather than assuming it holds. A case that no longer fits is usually still in use.

The cases are a reusable asset; the recurring costs are shelf space, till time and losses from the return loop.

A case occupies significantly more facing than the bare product, which reduces the range that fits on a fixture - a real merchandising cost that is rarely counted.

Every sale needs the case unlocked and emptied, which adds seconds and a physical action at the till.

Cases pile up at tills and disappear unless there is a routine to collect, count and return them.

Damage and loss mean an annual replacement rate, and a sudden rise is itself a signal.

Weigh all of that against the shrinkage on the protected lines. On lines that are not actually being targeted, the case costs more than it saves.

Counted, held securely once the store is closed and treated like till keys - a safer key outside the store opens every matching case in it.

Keys are typically common across a store or chain, which makes a single lost key significant.

Include them in the same handover count as the other security keys, and lock them away at close.

Keys coded to a single store contain the loss, and are worth requesting at purchase.

Keep a register of how many exist and who holds them, and treat a missing key as an incident rather than a re-order.

Magnetic-release designs behave like tag detachers: the release tool is the key and needs fixing down at the till.

Yes - most carry an EAS element or accept a label, so an item removed in its case still alarms at the exit.

That layering is what makes them effective: hard to conceal, and detected if carried out anyway.

The element has to match the store's technology, so check compatibility before ordering.

The till process then has two steps - unlock the case and, if the product itself is also labelled, deactivate that label.

Where the case carries the only EAS element, the product handed to the customer is unprotected, which matters for returns and for goods carried between store areas.

Decide and document which element belongs to the case and which to the product.

Small, high-value goods with a resale market, chosen from loss data rather than assumption.

Razor cartridges, memory cards, cosmetics, ink cartridges and over-the-counter medicines are the classic examples.

The common factor is high value in a pocket-sized package.

Over-protecting is a genuine cost: shelf space, till time and a slower shopping experience on lines nobody is taking.

Under-protecting is visible in the shrinkage figures, so the list should be reviewed against them regularly.

Be willing to remove protection as well as add it. Targeted lines change, and safers left on yesterday's problem are cases not available for today's.

With a container at each till, a scheduled collection, and a count - otherwise the shelf runs out while cases accumulate under the counter.

This is the same discipline hard tags need, and it fails in the same way.

A named role and a set time for collection is all it takes, and it should survive staff changes.

Counting on return pins a shortfall to a shift rather than to a quarter.

Damaged cases - cracked, hinge broken, lock worn - should be withdrawn rather than returned to the shelf.

Where a line is protected but the shelf has no cases, the stock goes out unprotected. That gap is invisible unless somebody is checking.