Competitor website change detection: seeing a price move the same day

It is 9:40 on a Tuesday morning. A rep reads a quote one last time before sending it: the same tier the buyer agreed to three weeks ago, plus the discount approved on the last call. The quote goes out. At 3:20 that afternoon a one-line reply lands: "The other vendor lists the same package for less per month, it's right there on their site." The rep opens the competitor's pricing page and finds a new table. The change went live the day before. Nobody at the company noticed, because opening that page every morning is not anyone's job.

This piece is about shrinking that gap to a single day. In order: what change detection actually compares, which pages earn a spot on the watch list and which are noise, where the line sits between following public information and breaking a site's rules, how to pick a check frequency, which movements deserve an alert and which deserve only a record, what to do in the first hours after a signal lands, how that signal becomes a pricing decision and a sales conversation, and the mistakes that quietly kill the whole setup.

What does change detection actually compare?

Change detection is not the vague activity of "keeping an eye on the competition." It is narrow and mechanical. You store what a specific public page looked like at a given moment, fetch the same page on the next check, and compare the two records line by line. What comes out is not an interpretation but a diff: which text disappeared, which number moved, which box appeared.

On a pricing page the places where a diff can mean something are finite, and you can write them down in advance: package names, monthly and annual amounts, the floor behind a "starting at" claim, the currency, the contract term, the seat or record limit inside each tier, the promotional banner across the top, and the checkmarks in the comparison grid. Almost everything else is cosmetic. Deciding what to look at before you look is also the first honest step in building a competitive intelligence practice that people trust.

A diff is not a decision. "The mid tier's monthly price dropped" is a fact; "they are cutting prices" is an interpretation, and it is often wrong. A feature may have quietly left that tier the same week, the lower number may be a ninety-day promotion, or a new structure may be pushing buyers toward annual billing where the effective price actually rose. The diff is raw material; the value comes from whoever reads it with the right questions.

Which competitor pages are worth watching, and which are noise?

A watch list works when it is short. A long one produces a pile of morning notifications nobody reads and gets muted within two weeks. Five to eight pages per competitor is plenty, and the useful ones are roughly the same in every industry.

• The pricing page: amounts, tiers, currency, contract length.
• The plan comparison grid: which capability moved to which tier, which usually shifts before the price does.
• Promotion and offer pages: temporary discounts tend to live at their own URL.
• Product and feature pages: a new module, a quietly removed capability, a reworded promise.
• The careers page: open roles tell you where a company is heading over the next two quarters earlier than the pricing page does.
• Release notes and announcements: the most honest record of what is actually shipping.

The noise side is just as clear: the blog feed, rotating testimonials, a footer carrying today's date, headlines that flip with an A/B test, image filenames. These report "changed" on every check and tell you nothing. Anchor the comparison to the meaningful block on the page rather than the whole document. Having sorted out your own packaging and tiering helps here, because you already know which row of their grid competes with which row of yours.

Where is the line between watching a public page and scraping data?

This distinction is legal and ethical rather than technical, and blurring it puts the company at real risk. Opening a public pricing page at reasonable intervals, the way any visitor with a browser would, and keeping a copy of what it said is ordinary competitive tracking; that page was published to be seen. What crosses the line is not looking at the page but working around the rules and protections the site has put in place.

Keep these boundaries firm: if the terms of use prohibit automated access, honour that; do not crawl paths disallowed in robots.txt; do not defeat a login, a paywall, a bot challenge or an IP block; do not create fake accounts to see gated pricing; and do not request pages at a rate that burdens someone else's server. Personal data stays out of scope entirely — names, contact details and profiles of a competitor's staff, customers or reviewers carry separate obligations under GDPR and equivalent privacy laws, and they have no place in a pricing watch list.

How you use the records follows the same line. Snapshots exist to inform your own decisions. Republishing a competitor's copy on your own site, dropping their screenshot into a campaign, or handing a prospect a distorted quotation is a different act with different consequences. This article is not a guide to scraping a competitor's website; it is about noticing, with discipline, when information anyone can already see has changed.

What do you gain by seeing a price change the same day?

The cost of the lag is countable. If their price moved on Monday and you learned about it on Friday, every quote sent across those four business days was written against a world that no longer existed. Some are still open, some are already lost, and the lost ones are not even tagged as price losses in the CRM, because nobody knew why they went.

The real gain of same-day awareness is not speed; it is the order of events. When the information reaches you before the buyer does, you walk into the call prepared. When it reaches you after, you are defending. The identical sentence produces two different outcomes.

None of this means responding to every move. Continuously repricing against demand, inventory and competitors does not suit every business model and often backfires in enterprise sales. Early information is valuable because it lets you decide calmly whether to respond at all, not because it lets you respond faster.

How do you choose a check frequency?

Polling every page hourly is both unnecessary and impolite to someone else's server. The right cadence depends on two things: how often the page genuinely moves, and how quickly a change there matters to you. If a pricing page changes three times a year, a daily check is generous. If a promotion page changes five times a month, daily may already be late.

A simple rule holds up well: look at how many times the page changed in the last ninety days, and check about five times more often than its average interval between changes. A page that moves monthly gets a weekly check; a page that moves weekly gets a daily one.

Managing that dial by hand is tedious and gets forgotten. The Refresh Frequency behaviour in competitor intelligence takes the decision off your desk: active competitors are collected more often, quiet ones less often, and you can still refresh any of them manually. Cadence becomes a moving range tied to how the competitor behaves rather than a number someone set once. Automated Competitor Monitoring runs that sweep across ad libraries, search, trends and map sources together, so what reaches you is the change rather than the chore.

Which changes deserve an alert, and which deserve only a record?

A monitoring system dies not the first time it sends a pointless notification but the third, so catching a diff and raising an alert have to be separate steps. Every diff is stored; only a diff past a threshold interrupts someone's day. The thresholds below are illustrative — tune them to your market.

Detected change	Example threshold	Action
Movement in a plan price	3 percent or more	Same-day alert to marketing and sales
Promotional banner added or removed	Always	Alert; record the end date
Feature moved between tiers	Always	Alert; route to the product team
New plan or tier introduced	Always	Alert; revisit positioning
Headline, image or ordering edits	No threshold	Recorded only
Blog, testimonial and footer copy	Ignored	Not watched

Who sets the threshold matters as much as the number. In the product this lives in Alert Thresholds and Weights at the organisation level; the comparison engine passes only threshold-crossing movement into the Change Detection and Alerts flow, and the in-app notification reaches members with marketing permission rather than everyone. Setting thresholds as low as possible is not diligence — it hands the filtering job to whoever opens the notification.

If you want an alert to trigger an automated move, extra care is needed. Anyone who has watched two automations fight over the same budget understands why wiring a competitor price signal straight into a bidding rule is risky. Let the signal come from the machine and the decision from a person.

What happens in the first hours after a change is caught?

When the alert lands, the first instinct should not be to call a pricing meeting. Start by verifying: open the page and confirm the diff is really there. Then measure the scope — one plan, the whole grid, or only the annual option? Then look for a clock: an end date, the word "limited", a condition in small print?

The third step is checking whether the change touches live business. Pull the open opportunities where that competitor is named; the ones with a quote sent but nothing signed are today's problem. If you know how a quote is constructed and what it commits you to, you can tell quickly which ones still sit inside their validity window.

The fourth step is writing it down: the date, the scope, and the decision taken that day. That record is worth far more three months later, because a competitor's pricing rhythm only becomes visible in hindsight. Historical Snapshots do exactly this — every collection is stored, so you can pick two dates, replay the difference row by row, and show in a meeting precisely when something moved. The AI Competitor Summary then compresses each collection into one readable paragraph.

How does a caught signal turn into a pricing decision?

A competitor lowering their price is not a reason to lower yours; at most it is an input. The questions come in order. Is the change permanent or temporary? Does the identically named plan really do the same job? In the deals you lost, was price genuinely the first objection, or the stated objection covering something else?

Price is one lever among several. Looking at the balance of product, price, place and promotion, the answer to a competitor's price move is often not price at all: implementation support, delivery time, contract flexibility, or moving one capability down a tier costs less and is harder to copy.

Even when you do respond with price, you do not have to change the list price to do it. Account-specific price lists and a discount matrix let you answer only in the segment that is genuinely under threat. Cutting the public number is the hardest decision to reverse, and it usually outlives the competitor move that caused it.

How does the sales team use this in a live conversation?

Marketing notices the change; sales pays for it. So the signal should not come to rest in a report but on the page a rep opens right before a call: one page per competitor, kept current.

The most common failure of competitive battle cards is that they are written once and left untouched for a year. A card with a wrong price line is worse than no card at all: the rep states a stale number with confidence and loses credibility in a single sentence. This is where change detection meets the card — when the page moves, the row that needs updating is already identified.

Use field data to decide what belongs on the card. If you measure which objections come up and how often in calls, you also know the context in which that competitor's name gets raised: price, integrations, or time to go live. The Win/Loss Against a Competitor tagging surfaces the same pattern from the other side, so you can say which rival actually costs you the most from records rather than from memory.

Which mistakes make the whole system useless?

The first mistake is scope. Put thirty competitors and forty pages under watch and the system works technically but not humanly; nobody reads the feed. Start with two or three real rivals and grow the list only as far as it keeps getting read.

The second mistake is ownerless alerts. A notification addressed to "the marketing team" is addressed to nobody. Give every competitor a named owner who leaves a two-line note about what they did after each alert. Three months on, those notes are worth more than the data itself.

The third mistake is watching only the other side. Your pages change too, and most companies know less about when their own pricing page moved than about their competitor's. That is what the Your Brand vs Competitor view is for: you measure your own brand against the same sources and read the scores on one chart, which settles whether the movement came from them or from you.

The fourth mistake is treating monitoring as an end in itself. A diff that never attaches to a decision is an archive. And sometimes the decision is to raise your own price, in which case the real work is announcing the increase without losing the customer — a communication problem, not a detection one.

If you want competitor page monitoring, alert thresholds and change history in one panel, you can open a free Rocketly account and start by adding your first competitor.

FAQ

Is monitoring a competitor's pricing page legal?

Opening a publicly published page at a reasonable cadence, the way an ordinary visitor would, and noting what it says is a normal part of competitive tracking. The line is crossed when you work around terms of use, robots.txt rules, or technical protections such as logins and paywalls. Personal data stays out, and stored content is not republished.

How often is often enough?

Match the page's real rhythm. A pricing page that changes a few times a year is well served by a daily check; promotion pages may warrant more. A fixed, high-frequency schedule burdens the other side's server and adds noise rather than information.

How do you tune alerts so the team is not overwhelmed?

Separate detection from alerting: record every diff, but notify only on diffs past a threshold. A percentage threshold on price, unconditional alerts on plan and grid structure changes, and record-only handling of cosmetic edits is a sound starting point.

Should you always answer a competitor's move with price?

No. First establish whether it is permanent or temporary and whether the identically named plan really does the same job. In most cases an answer built on delivery time, support quality or contract flexibility costs less and is harder for them to copy.