ONLINE PRICE MONITORING
Definition
A competing move detected and responded to within 3 hours
E-commerce retailer — monitors 8 competitors and 15,000 SKUs
between the time a price drop was detected at a competitor (2:00 p.m.) and the price was updated on the website (5:00 p.m.), compared to several days without automated monitoring.
A lower competitor price was detected at 2 p.m. for a television
Prices are updated and posted online at 5:00 p.m., after approval by the category manager
Why it matters
- Detect competitor movements in near real-time: a price change occurring in the morning can be identified and replicated within the same day.
- Cover a broad catalog: online monitoring allows you to track tens of thousands of SKUs simultaneously, whereas field auditing is limited to a few hundred.
- Feed the algorithms: monitoring data streams fuel dynamic and predictive pricing models.
Real-world example
An electronics retailer sets up price monitoring across 8 competitors and 15,000 references, with updates every 4 hours
The system detects that a competitor lowered the price of a television from €749 to €699 at 2:00 PM
At 4:00 PM, the algorithm proposes a price match of €699 to the category manager, who approves it
The price is updated on the website at 5:00 PM, exactly 3 hours after the competitor's price change
Without this monitoring system, the delay would have been several days.
How to measure and use it
Online price monitoring combines four components: 1) crawling competitor websites (handling CAPTCHAs, IP bans, and HTML structure changes), 2) product matching (EAN, text similarity, image), 3) data consolidation and cleansing, and 4) distribution to pricing tools
Enterprise solutions integrate these modules with freshness and coverage SLAs, while managing the scalability of monitored websites.
Common pitfalls
- Underestimating technical complexity: maintaining an up-to-date crawler across dozens of sites is a full-time discipline.
- Neglecting product matching: monitoring data with a 30% error rate in product matching is unusable.
- Confusing listed price with paid price: some websites employ personalized dynamic pricing, meaning the price displayed to a crawler may differ from the actual transaction price.
This topic is discussed in greater detail in our article on reliable product matching.
Mini-FAQ
Cette veille s'appuie sur un matching produit fiabilisé, condition indispensable pour comparer des références réellement équivalentes. Une fois structurées par catégorie dans un tableau de bord récurrent, ces données brutes alimentent le benchmark concurrentiel.
You might also be interested in these articles
Effective pricing management requires the rigorous integration of internal/endogenous data (costs, historical data) and external/exogenous data (competition, demand). This essential hybridization secures margins and objectifies trade-offs against market fluctuations. By structuring these signals, the organization transforms raw data into an operational profitability lever, deployable in practice in less than sixty days.
A price monitoring system that works perfectly in a pilot program does not deliver the same results on a large scale—not because the technology itself changes, but because the scale changes the nature of the problems to be solved: approximate matching, unmanaged alerts, and poorly calibrated data freshness become apparent where human oversight had masked them in a small scope.
Only 8% of companies actually manage to move their analytics initiatives beyond the pilot phase and deploy them organization-wide (McKinsey)—a reminder that the scaling of a price monitoring system depends primarily on methodology and governance, not on technology alone.
In-store surveys, web scraping, and retailer panels each answer a different question: what the customer sees on the shelf, what is displayed online at a given moment, and what has actually been sold. Confusing them is like answering the wrong question with the right data.
Among large multichannel retailers, the proportion of prices that change each month rose from 15% to nearly 30% between 2008 and 2017—a pace that a weekly survey or monthly panel is structurally unable to keep up with (Alberto Cavallo, NBER).