Imagine a typical morning in a store. A manager receives a list of twenty items with updated prices. A store associate has to find each product, print a new price tag, remove the old one, and put up the new one. Add to that a range of other important tasks competing for their attention. Physically, it’s hard for a person to keep up with all of this and update price tags on time.

Meanwhile, customers are already in the store. They choose products based on the prices they see on the shelf — only to discover a different price at checkout. Customers perceive this as deception and become frustrated.

Now multiply this situation across a chain of 50 stores with thousands or tens of thousands of SKUs, and imagine how energy- and time-consuming it is.

The traditional approach relies on a person walking through the store with a tablet or notebook, comparing price tags against reference data. It sounds simple, but in reality, it has many drawbacks.

First, the scale. A typical supermarket carries 3,000–5,000 products. Even if a checker could verify one price tag per minute (which is unrealistically fast), it would take an entire workday just to complete a single round.

Then there’s the human factor. Employees get tired, lose focus, rush to finish faster, and end up making even more mistakes.
If an issue is found, the employee must record it, find the responsible person, and follow up on the fix.

Manual checks also lack visibility. There’s no reliable analytics or clear picture of what’s happening in each store. Decisions end up being made based on subjective impressions. With store-level data across the network, management could make targeted, effective decisions where they are actually needed.

With automated price tag monitoring, you always know what’s happening on store shelves. This isn’t just about product availability and movement, but also about the information associated with them.

The system analyzes images of the sales floor, matches products with price tags and database records, and automatically detects discrepancies. Instead of walking the entire store and manually identifying such errors, you receive this information automatically. Staff only need to correct price tags or adjust promo displays where the system has detected inconsistencies with the plan.

As a result, you always have accurate and objective analytics at hand, enabling you to make more effective business decisions.

Ultimately, fines and customer complaints decrease, while customer satisfaction and revenue grow.

A price tag monitoring system operates in several sequential stages. Each stage solves a specific task and passes the result along the chain:

1. Cameras are installed in the store regularly capture images of the shelves.
2. The system processes these images: it detects products and price tags, matches them with database records, and identifies their exact names and SKUs.
3. Product-to-price tag matching. The system determines which price tag belongs to which product. This is not always straightforward, as price tags can shift and products may be moved.
4. Price validation — comparing the information on the price tag with data from the inventory system.
5. If the system detects any errors or discrepancies, it immediately sends an alert.

This is the overall workflow, though individual stages can be adapted to meet specific user needs.

Start with a pilot project
Don’t try to automate the entire network at once. Launch the system in a few stores or within a single product category. The first weeks will show where the algorithms perform well and where improvements are needed. Collect error cases and use them to retrain the model.

Send complex cases for human review
Sometimes the system is uncertain: poor image quality, unusual price tag formats, or ambiguous situations. In such cases, results should be routed for manual verification. This will give you valuable feedback for model retraining.

Keep price and promo data up to date
Promotions, temporary discounts, and new products appear every week. If reference data isn’t updated regularly, the system will start flagging errors that don’t actually exist.

Explain to employees that automation is a tool, not a punishment
Employees should understand that the system is there to help. It takes over the repetitive task of checking thousands of price tags and provides clear, actionable tasks that can be completed quickly.