POSM compliance: the right material on the right product

A refrigerator is on display. Its feature card is attached, its brand is visible, and the installation photo looks tidy. The job could still be wrong. The card might describe a different model, or a strip intended for an internal drawer might be stuck to the outside of the door.

That is the harder part of POSM compliance: verifying the relationship between a product, a piece of promotional material, and its intended location. Finding a logo in a photo does not establish that the installation matches the brief.

For appliance and electronics brands, this matters wherever similar products share a showroom. It also applies to branded coolers, demonstration units, and other equipment with model-specific graphics. Technology can help check these relationships, provided the standard is explicit and the photographs show enough to make a decision.

What a POSM compliance check needs to establish

POSM means point-of-sale materials: the feature cards, stickers, strips, panels, and other communication placed around or on a product. A basic availability check asks whether material is present. A placement check asks whether the correct version is installed on the correct product and surface.

Retail image-recognition systems already distinguish measures such as POSM availability, visibility, and planogram compliance. Vispera’s product overview lists these as separate retail execution measures. Model-specific placement adds a further requirement: a visible piece of material must be matched to an approved product-and-location rule.

Consider three installations:

Installation	What a presence check sees	What the brand needs to know
Correct feature panel on a lookalike model	A branded panel is present	Does that panel belong to this exact model?
Correct drawer strip on the outer door	The expected artwork is visible	Is it attached to the required interior surface?
Closed product with a tidy exterior	The outside looks complete	Were the required interior materials installed at all?

The third case is especially important. A missing view is not evidence of a missing installation. It is an unanswered question.

Why brands need more than an installation count

An incorrect feature card can describe a function the displayed model does not have. A misplaced panel can obscure the feature it should explain. A complete-looking exterior can hide an unfinished interior. Counting these as completed installations gives the brand a misleading account of its rollout.

The operational cost continues after the initial mistake. A reviewer has to find the relevant brief, identify the product, explain the correction, and check the replacement photo. If the installation is reported only as “non-compliant,” the field team may need another exchange before it understands what to fix.

The useful result is specific: which product, which material, which requirement, and which corrective action. That is also the difference between an image archive and evidence the account team can use.

Turn the placement guide into a reference the system can compare

A PDF guide is a useful starting point, but it often leaves knowledge implicit. A person may know that two similar strips belong to different model families, or that a retailer’s price card is outside the brand’s scope. Those distinctions need to become explicit rules.

Build a reference record for each approved model and material combination:

Reference field	What to define
Product identity	Model identifier, catalogue mapping, and distinguishing views
Material identity	Approved artwork, material ID, revision, and applicable campaign
Installation surface	Exterior door, interior drawer front, side panel, or another named surface
Placement zone	Position relative to product features, plus acceptable orientation and tolerance
Required evidence	The views and readable details needed to assess the requirement
Exceptions	Optional pieces, retailer-owned labels, and explicitly permitted substitutions

Use relative placement zones where they answer the business question. “On the upper drawer front” is often more useful than a pixel coordinate that changes with camera angle. Exact physical measurements need a calibrated capture method or another reliable scale reference; an ordinary perspective photo should not be treated as a ruler.

Model identity deserves its own evidence. A readable product label or catalogue-linked barcode can help distinguish visually similar units. GS1 defines the GTIN as an identifier for trade items; your catalogue still needs to map that identifier to the model and the correct artwork. A store price card is supporting evidence, but it can itself be attached to the wrong product.

Design the photo request around the questions

One wide photo is useful for context and often insufficient for detail. A practical capture brief can ask for:

• A complete-product view showing the appliance, its exterior materials, and its immediate display context.
• An open-interior view when any required material belongs inside. Ask for the actual drawer or shelf area, not simply an open door.
• A material close-up when small text or artwork differences distinguish approved variants.
• An identifier view showing the product’s model label or another agreed identifier.

Only request views that serve a defined check. Keep them linked to the same product and visit so a close-up cannot be silently combined with an unrelated overview. Glare, reflections, cropped edges, and people blocking the product are capture problems to handle explicitly.

Store and visit identifiers provide context. Timestamps and location metadata can support a check, but do not independently prove the photo’s origin. Duplicate screening can flag reused evidence; it cannot establish the truth of every new photograph.

Return a result for each requirement

A useful automated workflow separates product identification, artwork matching, and spatial placement. It then evaluates the observed evidence against the relevant reference version.

Each requirement should have three possible assessment states: pass, fail, or not assessable. Keep these separate from the overall workflow decision to approve, flag for review, or block a submission.

Here is an illustrative result for a fictional model, COOL-400:

Check	Evidence	Result
Product model	Readable model label matches the assigned product	Pass
Feature strip version	Visible strip matches the approved artwork	Pass
Installation surface	Strip is on the outer door; the brief requires an internal drawer	Fail
Interior completeness	Required interior view was not submitted	Not assessable

The next action is precise: move the strip to the specified drawer and provide the interior view. The workflow should preserve the original evidence and connect the correction to it, then recheck against the same brief version.

That record can feed an existing field-service tool, campaign dashboard, or agency review queue through an API. A reviewer should see the requirement, the evidence, and the reason together rather than reconstructing the decision from a generic score.

Validate the workflow before automating approval

Clean reference artwork tells a system what an installation should look like. It does not demonstrate how reliably it will work on store photos. Reflections, lookalike models, partial views, and small printed differences need testing with real evidence.

Start a pilot with a defined product family and an agreed checklist. Include correctly installed materials, wrong models, wrong artwork revisions, incorrect surfaces, and unanswerable photos. Have reviewers label the results against the same written standard and resolve disagreements before scoring the system.

Evaluate on submissions held apart from the examples used to configure the workflow. Avoid near-duplicate photos of the same installation appearing on both sides of that split. Measure:

• Incorrect installations that were approved, particularly model and surface mismatches.
• Correct installations that were unnecessarily flagged.
• Requirements that could not be assessed because the capture was insufficient.
• Reviewer time, correction completion, and the share of submissions that can safely receive an automatic decision.

Keep missing and unusable evidence visible in the reporting. Excluding it can make accuracy look better while hiding work the team still has to do. Expand by model family and check type when the evidence supports it, rather than treating one aggregate accuracy figure as permission to automate everything.

Where Steve fits

Steve’s POSM placement verification workflow is in early access. It applies the platform’s extraction, rules, evidence review, and delivery approach to a more specific question: does this material belong on this product, in this position?

This is verification against the brand’s installation standard. It does not certify product performance, prove the validity of a printed feature claim, or establish that the display stayed unchanged after the photograph.

For the broader rollout problem, see our guide to proving POSM installation from photos. For model-specific programmes, explore Steve for appliances and electronics, or bring your placement guide and representative photos to a pilot discussion. The first decision is which questions the evidence can answer reliably.