Most businesses that say they have pricing intelligence actually have price monitoring.
The difference sounds semantic. It is not. Monitoring tells you what competitors charge. Intelligence tells you what to do about it, and just as often, what to leave alone.
A team with the first and not the second ends up in one of two places. Either a daily report that nobody acts on, or an automated rule that follows every competitor move, including the ones that were mistakes.
This guide covers what pricing intelligence actually involves, the inputs that matter beyond the price itself, and how to build a working version without a data science team.
What pricing intelligence actually is
Pricing intelligence is the practice of collecting market pricing data, putting it in context, and turning it into pricing decisions.
It works in three layers, and each one depends on the one beneath it.
- Data. What comparable sellers charge, on which channels, and when that changed.
- Context. Whether each comparison is valid, and what else was true at the time: stock, promotions, delivery terms, who the seller is.
- Decision. What you change, what you deliberately ignore, and who is accountable for the call.
Most teams, and a good share of tools, stop at the first layer. Almost all of the value sits in the other two.
Monitoring versus intelligence
| Price monitoring | Pricing intelligence | |
| Question answered | What are competitors charging? | What should we charge, and why? |
| Output | A list of prices | A decision, or a reason not to make one |
| Cadence | As often as the data refreshes | As often as something meaningful changes |
| Main risk | Too much data, no action | Acting confidently on a bad comparison |
Neither column is optional. Intelligence without reliable monitoring underneath it is opinion. Monitoring without intelligence on top of it is a cost centre.
A worked example
Suppose a competitor cuts the price of your best-selling product by 12% overnight.
Monitoring reports the fact. That is the whole of what it can do.
Intelligence asks the questions that decide the response. Is it the same product, or a different bundle? Are they in stock, or clearing their last units? Is the new price tied to a promotion that ends on Sunday? Has the whole category moved, or only this item?
Each answer points somewhere different. A same-product, in-stock, category-wide move is a repositioning, and it deserves a response. A single-item cut on the last few units is stock clearance, and matching it gives away margin for nothing.
The data is identical in both cases. The right decision is the opposite.
The four questions it should answer
A working pricing intelligence process answers four questions on a regular cadence.
- Where am I priced out of the market?
- Where am I leaving margin on the table?
- What changed since the last review that actually matters?
- Why did it change?
The fourth question is the one monitoring alone never answers, and it is usually the one that decides whether you should react at all.
A competitor dropping a price because they are clearing discontinued stock calls for a very different response from a competitor repositioning their whole category.
The inputs that matter beyond price
Price is the most visible number and frequently the least informative on its own.
- Availability. A competitor who is cheaper but out of stock is not undercutting you. They are sending you customers.
- Promotions. A time-limited discount and a permanent repositioning look identical in a single snapshot. Only history tells them apart.
- Delivery terms. Customers compare the total landed cost, not the headline price. Free delivery on one side of the comparison changes the answer.
- Seller type. A brand’s own store, an authorised retailer and an anonymous marketplace seller are different kinds of competitor, and they deserve different responses.
- Marketplace position. On a marketplace, the seller holding the featured offer takes most of the sales. Being slightly cheaper than a seller who holds it may still leave you selling nothing.
- Match quality. The most important input, and the least visible one.
Why product matching decides everything
Every comparison assumes the two listings describe the same product. In practice that assumption fails constantly.
The same item appears under different titles, in different bundles and variants, and often without a shared identifier. One retailer sells the single unit, another the two-pack, a third bundles an accessory.
A comparison against the wrong product is worse than no comparison at all, because it looks authoritative and it gets acted on.
Matching products reliably across retailers, without shared identifiers, is where most in-house efforts stall. It is also the problem that price monitoring platforms such as Altosight were built to solve, since everything above it depends on getting this layer right.
Retailers and brands use it differently
The same underlying data serves two quite different purposes.
- For retailers, pricing intelligence is about position and margin: where you sit against the competitors your customers actually compare, and where a price could rise without losing the sale.
- For brands, it is about channel health: whether resellers respect advertised price policies, which sellers are undercutting authorised partners, and whether the market price of a product is quietly drifting down.
The inputs overlap almost entirely. The questions do not, which is why a setup designed only for one side often serves the other poorly.
How to build it without a data science team
Pricing intelligence does not require a modelling function. It requires a few decisions made in advance and someone who owns them.
- Define the competitive set. Three to eight competitors that your customers genuinely cross-shop. Not everyone who sells the category.
- Segment the catalog. Traffic drivers that customers compare, margin makers where you are differentiated, and a long tail nobody price-checks. Each group deserves a different level of attention.
- Write the rules before looking at the data. Price floors, ceilings, and the size of movement that triggers a review. Rules written in the middle of a price war are rarely good rules.
- Assign an owner. Insight without authority changes nothing. Someone needs to be accountable for acting, and for deciding not to.
- Review on a cadence. Weekly for the products customers compare, less often for the rest. Continuous watching produces anxiety, not decisions.
What a useful weekly output looks like
The most common reason pricing intelligence fails is that the output is designed for completeness rather than action.
A useful weekly view is built around exceptions, not totals:
- Products now priced outside your target range against the competitive set
- Products where you could raise the price and remain competitive
- Competitor moves large enough to cross your review threshold, with the likely reason attached
- Comparisons flagged as unreliable, so nobody acts on them by mistake
Everything that stayed within its expected range can stay off the page. If the report is long, it is not being read.
Where it goes wrong
- Drowning in data. More competitors, more products and more alerts feel like progress. Past a point they produce paralysis, and the report stops being read.
- Reacting to noise. A meaningful share of competitor price drops are stock clearance, feed errors or plain mistakes. Automated matching transmits those mistakes across a whole market within hours.
- Nobody owns the decision. The data arrives, everyone sees it, and no one is accountable for acting. This is the most common failure, and the least technical.
- Treating cheapest as the goal. Availability, delivery speed and trust often outweigh a small price gap. Racing to the bottom is a strategy for pure commodities, not a default.
Evaluating pricing intelligence tools
When you compare pricing intelligence tools, the feature lists tend to look alike. The differences that matter sit underneath them.
- Matching accuracy, and how it is verified. Ask how products are matched when no shared identifier exists, and who checks the result.
- Coverage. Independent webshops as well as marketplaces, and every country and currency you actually sell in.
- Context, not just price. Stock, promotions, delivery terms and seller identity, captured with dated history.
- The path to a decision. Alerts, rules, exports or an API. Data that cannot reach the person who decides is decoration.
- Freshness you can see. Every price should show when it was last checked. A comparison against a week-old number is a guess wearing a timestamp.
- Time to value. How long until the first trustworthy report, and how much of your team’s time the setup consumes.
Where to start
Pick one category, one competitive set and one month. Answer the four questions for that category only, every week, with a named owner.
At the end of the month you will know whether your data is trustworthy, whether your rules make sense, and where the real gaps are. That is a far better basis for scaling than a catalog-wide rollout on day one.
The goal is not more data. It is fewer, better decisions, made faster, by someone who owns them.
