What is a pricing alert - and how should retailers use them?

A pricing alert is a simple thing with a hard design problem attached. The simple part: it's a notification that a condition you care about has occurred - a competitor undercut your KVI, a price index crossed a threshold, a MAP violation appeared. The hard part: the difference between alerts that drive decisions and alerts that get ignored is entirely in how you configure them. Alerting is the human-attention layer of price monitoring, and most teams get it wrong in the same predictable way. Here's how to get it right.

What an alert is actually for

An alert exists for exactly one purpose: to summon human judgment to a situation that automation shouldn't handle alone. That framing clarifies everything. If a situation can be handled by a rule within guardrails - competitor moved, policy says match, floor says how far - it shouldn't generate an alert at all; it should just execute, and appear in the log. Alerts are for the exceptions: the MAP violation that needs your brand team, the unusual gap that suggests a match error, the index drift that signals strategy rather than tactics.

The most common failure is inverting this - alerting on everything and automating nothing. A feed that pings on every competitor price change trains your team to ignore it within a fortnight, and an ignored alert system is worse than none, because it provides false confidence that someone is watching.

The five alerts worth configuring

1. KVI threshold breaches. Your key value items - the products that shape price perception - drifting beyond a set gap versus tracked competitors (2-3% is a sensible start). These deserve same-day attention because hours matter on them.

2. MAP violations. A competitor advertising below minimum advertised price. This alert routes to brand relations, never to auto-matching - responding in kind would put you in breach too.

3. Suspected match errors. A "competitor price" that suddenly moves 40% is more often a matching problem than a market event - the listing changed, the bundle changed, the match decayed. Flagging improbable moves for verification protects every downstream decision, since stale and mismatched competitor data corrupts quietly.

4. Index drift. Your competitive price index for a category crossing a defined band - say, above 104 or below 96. Item alerts catch fast moves; index alerts catch the slow ones nobody notices day-to-day. Reading these correctly is its own skill, which our step-by-step CPI guide covers.

5. Guardrail collisions. A rule repeatedly wanting to price below a margin floor is telling you something - usually that a competitor's sustained position or a cost change has made your current strategy untenable on that SKU. That's a strategy conversation, not a price change.

Design rules that keep alerts alive

Tier by product importance. Tight thresholds on KVIs, loose or none on the long tail - attention budgeted where perception is formed. Digest what isn't urgent. One morning summary for semi-important movements beats forty pings. Attach context to every alert: the competitor, the gap, the margin room, the recommended response - an alert that requires research before action is half an alert. Review thresholds quarterly, because a threshold set in January describes January's market.

From alert to action

The final measure of an alerting setup is the gap between notification and decision. Alerts feeding a four-day manual cycle just document your losses in real time; alerts landing in a workspace where the recommendation, reasoning, and guardrails are already attached close in minutes. That's how alerting works inside Retailgrid's agentic pricing - routine moves execute, exceptions arrive as decisions-in-waiting.

See the exception queue live on a real dataset in the interactive demo - no signup needed.