HOURLY
Competitor pricing pages
Price, promo badge, shipping threshold, stock. The value of the alert decays fast: knowing on Friday that a rival cut prices on Monday is trivia. Knowing within the hour is a decision.
Website change monitoring means re-reading a page on a schedule and telling you when something you care about is different. WebRobot watches fields, not pixels: you name the price, the stock status, the job title or the clause, and the robot alerts Slack, email or a webhook the moment that value moves. A rotating hero banner will not wake anyone up. A competitor cutting a price by 20 percent will.
Last updated July 2026
Robot console · WR-01
Standing by Running · s Complete · rows
1 · Pick a target
3 · Fields to extract
Agent log
Crawl graph
Extracted data · rows
Want this data fresh every morning, without lifting a finger?
| Method | What it compares | Fires on | Fails when |
|---|---|---|---|
| Visual diff | Screenshots, pixel regions | Any visible movement in the region | Carousels, ads, cookie banners, A/B tests |
| Text or HTML diff | The raw page source or the text blob | Any character difference | Timestamps, session IDs, view counters |
| Field-level | Named values: price, stock, title, date, clause | Only the value you asked about | Rarely: intent survives redesigns |
Visual and text diff tools are genuinely useful, cheap and often free, and if you are watching one public page for "did anything at all happen", they are the right pick. They fall apart at scale. Watch 40 competitor product pages with a pixel diff and you will get several hundred alerts a week, almost all of them a rotating promo image. The team stops reading them by Thursday, which means the one alert that mattered gets ignored too.
Field-level monitoring inverts that. The robot extracts the same structured row every run ("SKU, price, promo flag, stock"), diffs the row against the last one, and only fires when a named field is different. The alert carries the old value and the new one, so it is readable without opening the page.
Paste a list of URLs, or point the robot at a category page and let it paginate to find the products itself. One robot can hold the whole watch list. If the pages sit behind a partner login, store the credentials once and the agent signs in on every check.
"Price, promo badge, stock status." Everything else on the page is explicitly not a change: the reviews counter can tick, the banner can rotate, the "12 people are viewing this" widget can do whatever it likes. Nothing fires.
Daily on Launch, hourly with change alerts on Scale, real-time monitors on Autopilot. Frequency should track how fast the value moves and how much a late answer costs you. Hourly is the right default for competitor pricing.
Alert on any stock flip, but only on price moves above two percent. Ignore currency rounding. Batch low-priority changes into one digest and send the urgent ones on their own. This is the step that decides whether the channel survives its first month.
Slack for the analyst who reprices, a webhook into the pricing engine for automatic response, Google Sheets for the running history, the REST API for the warehouse. Alerts that land in an inbox nobody owns are the same as no alerts.
The full walkthrough for the pricing case is in how to monitor competitor prices, and the dedicated price monitoring tool page covers repricing workflows in more depth.
HOURLY
Price, promo badge, shipping threshold, stock. The value of the alert decays fast: knowing on Friday that a rival cut prices on Monday is trivia. Knowing within the hour is a decision.
HOURLY
New postings are a hiring signal and a sales signal. Watch title, team, location and posted date. A company opening six infrastructure roles is telling you something about its roadmap.
REAL TIME
Stock flags, allocation windows, ticket availability. This is the one case where minutes matter and a daily check is worthless. Real-time monitors are on the Autopilot plan.
DAILY
A supplier quietly editing a liability clause or a platform changing its API terms. Watch the clause text and keep the previous version, so legal can see exactly which sentence moved.
DAILY
Agency notice boards, tender lists, license registers. Mostly quiet, occasionally decisive, and almost never equipped with an RSS feed or an API.
DAILY
Plan names, feature lists, tier limits. Positioning changes show up here weeks before anyone announces them. Pair it with a data extraction tool run to keep the full history.
| Channel | Carries | Best for | Plan |
|---|---|---|---|
| Slack message | Field name, old value, new value, URL, timestamp | Analysts and buyers who act same day | Scale and up |
| Email digest | Batched list of changed rows | Legal and policy watching | All plans |
| Webhook | JSON payload with the diff | Automatic repricing or ticket creation | Scale and up |
| Zapier | Trigger with the changed row | Routing into Asana, Jira, HubSpot | Scale and up |
| Google Sheets row | Append-only change history | Trend charts and audit trail | All plans |
| REST API | Latest dataset and per-run diffs | Feeding a warehouse or dashboard | All plans |
Change alerts and hourly checks begin on Scale at $249 per month ($199 billed yearly), which also includes agent actions, so monitored pages can sit behind a login. Launch at $79 covers daily checks on 5 robots and 10,000 records. Autopilot at $699 adds real-time monitors and 1,000,000 records. See pricing for the full breakdown, and security for how stored credentials are handled.
A selector-based monitor points at a CSS path. The site renames the class, the path resolves to nothing, and the monitor reports calm forever. WebRobot stores the intent ("the price next to the buy button") and re-locates the field after a layout change, then notes the repair in the run log.
Noise is the other silent failure. If the channel fires 200 times a week, it is functionally off. Thresholds, field scoping and digests are not conveniences, they are what keeps the signal credible.
A login expiring, a region block, a page moved to a new URL. Runs record status and row counts, so a run that returns zero rows where it returned 200 yesterday is reported as a failure, not as "no changes". More on run mechanics in how it works.
FINAL ASSEMBLY
Name the field, set the cadence, pick the channel. The robot watches the page and only speaks when it matters.