The Cost of Waiting Until Campaign End
A planogram is usually configured for a campaign window of four to eight weeks. In a manual process, compliance is often checked at the end of that window, when someone walks the store, compares the display with the intended layout, and records what has changed. By then, a misplaced product, an empty facing, or a shifted section may have been visible to shoppers for weeks. The cost is not mainly the absence of a planogram. It is the time gap between setting the standard and detecting drift from it.
Camera-based compliance measurement changes that cycle. Instead of treating the campaign-end walk as the first reliable measurement, the system checks the display during the same shift in which the change appears. The result is not a replacement for every form of store judgment. It is a more frequent measurement of one specific operational question: how closely does the observed display match the configured planogram?
At Youtiligent, that measurement is built around three elements: a reference planogram, mapped display zones, and a facing-based compliance score. The purpose of this article is to explain how those elements connect, what an exception report contains, and where the method has clear limits.
What the Reference Planogram Contains
A reference planogram is not simply an image of an attractive shelf. It is a configuration file that describes the expected structure of the display in a form the measurement system can compare with a live camera view.
The configuration contains three important types of information:
- Shelf section definitions: the sections or areas that make up the display, such as a top shelf, a promotional bay, or a defined group of shelves.
- Expected product facing positions and counts: the products expected in each section and the number of visible facings assigned to them.
- Zone boundaries: the spatial limits that tell the system where one measured area begins and another ends.
A facing is the visible front position of a product on the shelf. If a reference calls for three facings of a product in a particular section, the baseline expects those three visible positions to be present there. The baseline can therefore represent more than whether a product exists somewhere in the display. It represents where the product should appear and how many visible positions it should occupy.
This distinction matters operationally. A display can contain the correct products and still fail to follow the planogram if products have been moved into the wrong section, if one product has displaced another, or if the intended facing count has been reduced. The reference planogram gives the comparison a defined structure instead of asking a manager to make a general visual judgment.
Planograms are loaded once per campaign window, typically every four to eight weeks. That does not mean the display is measured only once. The planogram is the stable reference, while the camera view provides repeated observations against it during the campaign.
Mapping Zones in the Dashboard
The planogram comparison happens by zone. In the Youtiligent dashboard, a user defines zones by drawing the boundaries of the relevant shelf sections over the camera view. A zone might cover one shelf section, a promotional display, or another area that needs its own target and baseline.
Each zone receives two connected settings. The first is a fill rate target, which describes the expected level of product presence for that area. The second is the planogram baseline, which describes the expected product facings and positions. Together, these settings distinguish a display that is merely occupied from one that is arranged according to the campaign standard.
Zone mapping is a practical configuration task, not an abstract computer vision exercise. The boundary needs to follow the physical display visible from the camera. If a line cuts across two unrelated sections, the resulting score can combine conditions that should be investigated separately. If a zone is too narrow, it may exclude facings that belong to the section. If it is too wide, it may include neighboring products or fixtures that do not belong to the reference.
For that reason, zone design should follow the way the merchandising team assigns responsibility. If a manager needs to know which section drifted, the section should be represented as its own zone. A single score for an entire wall may show that something is wrong, but it will not provide the same operational direction as separate scores for the promotional bay, the upper shelf, and the adjacent category section.
There are also physical constraints. Zones can become difficult to interpret when sections overlap visually, when a shelf edge is hidden, or when the camera angle does not cover the full display. Low-light sections can reduce the reliability of what is visible in the image. These conditions do not make measurement impossible, but they should be handled during setup rather than treated as scoring anomalies after deployment.
How the Compliance Score Is Calculated
The basic scoring method is deliberately understandable. For each zone, the system compares the observed product facings in the live camera view with the expected product facings in the reference planogram.
The core relationship is:
Zone compliance ratio = observed product facings divided by expected product facings
The zone ratios are then averaged across the display to produce the compliance score. The score is expressed on a 0 to 100 scale. A display with the expected facings present in the expected sections will trend toward full compliance. Missing facings, displaced products, and section-level deviations reduce the ratio for the affected zone and therefore reduce the display score.
Consider a simplified display with three zones. The first has an expected count of ten facings and eight are observed. The second expects six and has six observed. The third expects four and has two observed. The individual ratios are 0.80, 1.00, and 0.50. The display-level result is based on the average of those zone ratios, represented on the 0 to 100 score scale. The important point is that the score is tied to observed and expected facings, not to an opaque pass or fail label.
This approach also explains why a display score alone is not enough. Two displays can have the same overall score while having different operational problems. One may have a small amount of drift across several zones. Another may have one severely disrupted section and several sections that remain compliant. The section-level comparison preserves the information needed to tell those cases apart.
The score is computed per shift. That timing is important because store conditions change throughout the day. Products can be moved during replenishment, promotional stock can be depleted, and a shelf can be corrected before the next shift. A shift-level score creates a record of when the deviation was present instead of reducing the campaign to one end-of-period observation.
The system does not need to claim that every visual detail is perfect to be useful. It needs to provide a consistent comparison within the defined zones, against a known reference, with enough detail to direct follow-up. The score is a measurement of planogram adherence, not a general assessment of store execution.
From a Low Score to an Actionable Exception
Youtiligent generates a daily score for each display. When the score falls below 80, the system automatically triggers an exception report to the merchandising contact. This threshold turns a continuous measurement into a work queue. The merchandising team does not need to review every display with the same urgency. It can focus first on displays that crossed the configured compliance boundary.
The report identifies which specific section drifted. That detail is the difference between an alert and an operational instruction. A message that says a display scored 74 leaves a manager searching the entire fixture. An exception report that identifies the lower-right promotional section gives the manager a defined area to inspect, correct, or route to the appropriate store contact.
A report can support several common follow-up questions:
- Which display fell below the compliance threshold?
- Which section or zone caused the exception?
- Was the issue associated with missing or changed facings against the reference?
- Did the deviation appear during the current shift or persist across shifts?
This changes the workload from scheduled inspection to exception handling. Floor managers are not asked to conduct a complete planogram walk simply to discover that most of the display is correct. They can investigate the reported section, correct what is within their control, and escalate the issue when the cause is replenishment, stock availability, or an incorrect campaign setup.
The distinction is especially useful for merchandising teams managing many locations. A weekly floor walk may provide a detailed snapshot, but it also consumes time at stores that may not need intervention. Daily exception reporting directs attention toward the displays with measurable drift and preserves a record of where the problem occurred.
What We Observed in a Pilot Cohort
A regional pharmacy group in our early-access pilot program used the method across a promotional display network. The planogram compliance score lifted from 71% to 89% within 28 days. The merchandising team moved from weekly floor walks to running exception reports and focusing follow-up on the sections named in those reports.
The result does not mean that camera measurement, by itself, corrected every shelf. The operational change was the feedback loop. The team received a shift-level signal, saw which section had drifted, and could direct attention before the campaign ended. The 18-point lift over the four-week pilot window illustrates the value of shortening the interval between deviation and response.
For implementation, the useful question is not whether a score can be produced. It is whether the score is connected to a planogram format, a workable zone map, a defined threshold, and a person who can act on the exception. Without those links, a number adds reporting but not necessarily better execution.
What the System Measures Well, and What Requires Judgment
Camera compliance scoring measures the relationship between visible product facings and the expected facings in the configured reference. It is well suited to identifying missing or displaced visible facings, comparing defined shelf sections, tracking score changes by shift, and directing a team to the zone that needs attention.
There are important boundaries:
- Product quality: the score does not determine whether packaging is damaged, expired, contaminated, or otherwise unsuitable for sale.
- Facing depth beyond the camera angle: the measurement concerns visible facings. It does not establish how much stock is positioned behind the front row when that depth cannot be seen.
- New product types: a product that is not represented in the reference planogram may require a planogram update or human review. The system cannot infer the intended campaign position for an unconfigured product.
- Coverage gaps: a camera that does not see the full display cannot support the same confidence as a view covering all relevant zones.
- Overlapping zones: boundaries that include the same physical area, or that do not match the shelf structure, can make the resulting ratios difficult to interpret.
- Low-light sections: poor lighting can reduce the visibility needed for a reliable comparison and should be corrected or flagged during setup.
These limits define the proper role of the score. It is a measurement layer for visible planogram compliance, not a substitute for product inspection, campaign governance, or merchandising judgment. The quality of the result depends on a reference that reflects the current campaign, zones that match the physical display, and camera coverage that includes the areas being scored.
If you want to see how the compliance scoring would work with your planogram format, send us an example. We can show you what zone mapping looks like before any commitment. [email protected].
Back to Store Ops Insights