Choosing Your Competitive Set: Which Stores Are Actually Worth Monitoring

Most monitoring programs that fail do not fail at the alert stage. They fail earlier, at the moment someone typed a list of store URLs into a tool without asking why each one was there. Every threshold, cooldown, and review ritual that follows operates on that list. If the list is wrong, the rest of the setup is tuning noise.

Operators usually diagnose the resulting fatigue as a threshold problem. The alerts are too sensitive, so the fix must be a higher percentage or a longer cooldown. Sometimes that is true. More often the alerts are working exactly as configured and the problem is that a third of the stores generating them were never worth watching in the first place.

This guide is about that earlier decision. The competitor monitoring checklist and the operator playbook both give you a tier structure and a rough count for each tier. Those are useful for how many. This is about which, and about who to leave out.

The Set Is the First Filter, Not the Last

It helps to think of monitoring as a pipeline. Stores go in one end, change events come out the other, and everything in between (scan frequency, alert types, thresholds, digests) is a filter. Filters can only remove. They cannot turn a low-quality input into a high-quality output.

A store in a different price tier will generate price drops that are technically real and completely irrelevant to your customer. A marketplace will generate hundreds of catalog changes a week that reflect seller churn rather than merchandising decisions. A dormant storefront will generate nothing for months and still occupy a slot in your weekly review. None of these are fixable downstream. The only fix is to not include the store.

The practical consequence is that time spent on selection pays off more than time spent on tuning. An hour deciding which six stores deserve attention will do more for your signal quality than a month of threshold adjustments on a list of twenty.

Four Criteria That Predict a Useful Store

A store earns a place in your set when it meets most of the following. None of them are absolute, but a candidate that fails three of the four almost never turns out to be worth the slot.

Category and price band overlap

Same industry is not the test. Same shelf is the test. The question to ask is whether a shopper comparing one of your top products would plausibly have the competitor's equivalent open in the next tab. If the answer is yes for a meaningful share of your revenue-driving products, the store passes.

Price band matters as much as category. A store selling the same product type at three times your price is not competing for the same purchase decision, and its price moves will not change what your customer does. Look for overlap in the band where most of your revenue sits, not at the edges of your range.

Comparable assortment size

A forty product catalog and a four thousand product catalog produce very different data even in the same category. In the small catalog, every new product and every removal is a deliberate decision that someone made. In the large one, the same events are routine inventory rotation and mean far less individually.

This is not a rule against monitoring larger stores. It is a warning that their events need to be read differently, and that a set mixing very different catalog sizes will make your weekly review harder to interpret. When you have a choice between two similar candidates, prefer the one whose assortment is within a few multiples of your own.

Shared customer intent

Two stores can sell the same category to customers with entirely different jobs to be done. A specialty running shoe retailer focused on fit and a fashion sneaker store carrying some of the same models are in the same category, in an overlapping price band, and are barely competitors. Their customers arrive with different questions and leave for different reasons.

Intent shows up in how a store merchandises: the collection names, what the homepage leads with, what the product descriptions emphasize, who the imagery is speaking to. Spend five minutes on a candidate's site as a shopper would and it is usually obvious whether you are solving the same problem for the same person.

The store actually changes

This is the criterion most operators skip, and it is the one that most directly determines whether the store will ever produce a useful alert. A storefront that has not changed a price or added a product in four months will not start doing so because you are watching it.

You can check this before committing. Sort a collection by newest and look at how recent the additions are. See whether the sale or clearance collection rotates or has held the same items for a season. Look at whether the catalog appears to be growing, holding steady, or quietly shrinking. Two weeks of baseline monitoring will confirm what the visual check suggested, and a candidate that generates nothing in that window should be parked rather than kept.

Who to Leave Out, and Why

Some categories of store fail the criteria so reliably that it is worth naming them explicitly.

Marketplaces. A marketplace storefront is a platform, not a merchandising decision. Prices are set by many independent sellers, catalog changes arrive as a constant stream of listings appearing and disappearing, and nothing about the aggregate reflects a strategy you could respond to. If a specific brand on a marketplace matters to you, watch that brand's own storefront instead, where the changes are decisions.

Giant generalists. Very large multi-category retailers change thousands of prices a day, most of them through automated repricing rules rather than human judgment. Any signal that exists is averaged away in the volume, and the events you receive will not tell you what the retailer is thinking because in most cases nobody was.

Stores in a different price tier. A luxury retailer is not a competitor for a value brand even in an identical category, and the reverse is equally true. Their price drops are not visible to your customer's decision, and including them will make your price alerts look busier than your actual market is.

Dormant storefronts. A store that has stopped moving costs review time and returns nothing. The one exception is a store that appears to be winding down, where a shrinking catalog is itself worth noting. Note it once and move on. It does not need a standing slot.

Brands you stock. Suppliers whose products you resell are worth watching, but for a different reason and in a different list. You are tracking their published pricing for MAP compliance and policy changes, not competing with their merchandising. Keep them separate so they do not distort your read of the competitive set.

How Many Stores Can You Genuinely Review?

The right number is not a store count. It is a time budget. Decide how many minutes per week you will actually spend on competitive review, then divide by how long a meaningful look at one store takes.

For an operator who knows their set well, a real review of one store's weekly changes takes five to eight minutes. That is enough to read the events, decide whether any of them matter, and note a response or a non-response. A forty five minute weekly budget therefore supports six to nine stores at genuine attention. Not twenty five.

Many teams hold a list of twenty five and review five. The other twenty are not being monitored in any useful sense. They are generating events nobody reads, which trains the team to treat the whole alert stream as background noise, which eventually costs them the five that mattered. A short, deliberate list beats a long one because a short list is the one you will actually read.

There is a reasonable case for alert coverage being wider than review coverage. A store can sit outside the weekly review and still carry one or two narrow alerts, such as a price drop on a specific overlapping product. But each of those stores still has to earn its place under the criteria above. Wide alert coverage on stores that fail the criteria is just noise with a longer delay.

Start with five. Add a sixth when you can name the decision you expect it to inform.

Reviewing and Pruning on a Cadence

A competitive set is a living list. Stores reposition, launch new lines, go quiet, or get acquired. Your own catalog and price band drift too. A set that was right in January is often wrong by July, and the drift is invisible unless you schedule a look at it.

Once a quarter, review the list with four questions per store:

• How many events did this store generate in the last ninety days?
• Did any of those events change a decision you made?
• Has the store's positioning drifted out of your price band or category?
• Did you actually look at it during weekly reviews, or did you skip past it?

A store that generated events but never changed a decision is a candidate for demotion to a narrower alert-only role. A store you consistently skipped is a candidate for removal regardless of what it generated, because your own behavior is telling you it does not matter. A store that drifted out of your band should be parked with a note explaining why, so you can revisit it if it drifts back.

Two habits make this work. First, swap rather than add. When a new candidate earns a place, ask which current store it is replacing, and only expand the list when you genuinely cannot answer. Second, keep a one line note for every store stating why it is in the set and when you last confirmed that. If you cannot write the line, the store does not belong. This is the same discipline the weekly operator review applies to individual events, applied to the list itself.

A Worked Selection

Consider a mid-price home fragrance brand: candles and diffusers, around 120 products, most revenue between 30 and 60 per item. The operator's first draft of a competitive set has fourteen stores on it, assembled from search results, a trade show, and a few brands that show up in their customers' social feeds.

Walking the list against the criteria removes most of it. Three stores are direct fits: same category, same band, same gifting and self-purchase intent, and all three added products in the last month. Two are luxury fragrance houses whose entry price is above this brand's ceiling, so they are parked with a note. One is a marketplace category page, removed. Two are very large home goods retailers whose candle sections are a rounding error in their catalogs, removed. Two have not added a product in over six months, parked as dormant with a date to recheck. One sells a gift box subscription that happens to include candles, which is a different customer intent, removed. One is a slightly larger adjacent store that sells home fragrance alongside bath products at a comparable band, kept as an adjacent watch. The remaining two are suppliers whose products the brand stocks, moved to a separate list for pricing policy checks.

The result is four stores under weekly review, three parked with reasons and revisit dates, and a supplier list handled separately. Every alert that arrives from this point is from a store that was chosen on purpose.

Where Bonesaw Fits

Bonesaw monitors publicly accessible storefront state: what is listed, what it costs, and whether it is available to buy. It does not see a competitor's internal inventory system, their order data, or their analytics, and neither does anything else that works from the outside. That boundary is also what makes the selection criteria above practical, because every one of them can be checked against data you can actually collect.

The most useful way to use that is as a trial period. Add a candidate, let it run for two or three weeks, and look at what it produced. A store that generated a handful of meaningful catalog and price changes has demonstrated that it moves. A store that generated nothing has answered the question for you. Deciding on evidence is easier than deciding on instinct, and removing a store that turned out to be quiet costs nothing.

If you are starting from scratch, the getting started guide walks through adding the first store. The criteria in this guide are what to bring to the second one.

Frequently Asked Questions

Is it better to start with too many stores and cut, or too few and add? Too few. A short list that you review every week builds the habit and the judgment that let you evaluate additions well. A long list that overwhelms you in the first month tends to get abandoned before any pruning happens. Five stores reviewed carefully will teach you more than twenty skimmed.

Should I include the market leader even if they are much larger than me? Only if they pass the criteria on their own merits. A leader in your exact category and band whose catalog changes are deliberate is a strong candidate regardless of size. A generalist that happens to lead in revenue while treating your category as a small section usually is not. Size is not the test.

What about stores that only compete on a few products? Give them a narrow role. A store with two or three overlapping products can carry a product-level price or availability alert without being part of the weekly store review. That keeps the coverage without spending review time on a catalog that is mostly irrelevant to you.

How do I know whether a store is dormant before I add it? Sort a collection by newest and check the recency of additions, look at whether sale collections rotate, and see whether the catalog is growing or shrinking. Then confirm with a couple of weeks of monitoring. A store that produces no events in that window has told you what you needed to know.

My competitors are on different platforms. Does that change the selection? Not the criteria. Category, band, intent, and change frequency apply regardless of what the store runs on. What differs is what can be observed on each platform, which the cross-platform monitoring strategy covers in detail.

How often should I revisit the set? Quarterly is a good default, with an extra pass before peak season when positioning tends to shift. If you notice yourself skipping the same store in three consecutive weekly reviews, do not wait for the quarter. That is the signal.

---

Bonesaw is a product of MoonsLink. Monitoring capabilities described in this guide reflect publicly accessible product and storefront data collected through standard web protocols. Bonesaw does not access private or authenticated data. All data collection respects robots.txt directives and site access policies.