Analysing the competition

Classic competitive analysis produces a twelve-column table nobody reads again. What follows fits into three questions, and each one leads to a decision you can make this week.

This is the fourth step of a full study. If you have not yet measured demand or collected prices, start with the six-step method: analysing competitors in a market whose size you do not know leads nowhere.

What is not worth doing

Let us start with what you can stop doing:

• Counting competitors. "There are forty-seven" says nothing. Forty-seven sellers where three hold most of the market, and forty-seven on equal footing, are opposite situations, and the count is identical.
• Copying out their "about" pages. That is communication, not data.
• Following their social accounts. Time-consuming, and follower counts do not tell you what they sell.
• Filling in a SWOT matrix. It arranges what you already know. It teaches you nothing you did not write into its boxes yourself.

Question 1: How is the market split?

One measure only: the top seller's share of the visible offer. It tells you who you are dealing with, which the total count never does.

When one player dominates, they set the reference price, and your positioning is defined relative to them: either you match it and compete on something other than price, or you take a position they cannot occupy: the highly specialised, the premium, the service.

When nobody dominates, the market is open. Good news for entering, less so for staying: nothing will stop the next entrant after you, using the same arguments.

MarketRaccoon draws the line at 30% of the collected offer for the top seller: above it the display reads "concentrated", below it "fragmented". That is our display threshold: it gives a stable reference from one analysis to the next, which is its only purpose.

Do not count single-product sellers

A seller appearing with one listing drifts in and out of the data at random. Tracking them means measuring noise. We only keep one once at least three of their products are found, precisely to avoid announcing movements that never happened.

Question 2: What do they actually sell for?

Not the price advertised on a homepage: the price actually charged on product pages, promotions included.

Three figures are enough (lowest, median, highest), then the distribution: how many products in each price band. It is the distribution that carries the information, not the extremes.

It reveals what no summary table shows: the empty bands. A cluster of products at the entry level, then a gap, then a few very high references. That gap is either a place to take or a graveyard. To find out which, look for someone who occupied it and left.

The trap in price collection

A price collection is not reproducible identically from one day to the next: results vary, sellers appear and disappear. We measured this on our own collections, and it is why our tool forbids itself certain comparisons.

Two precautions apply to you as well. Compare product by product: never two overall averages, which move as soon as the product list changes. And only conclude there is a rise or a fall if the gap clearly exceeds the variation you see between two collections made on the same day. Below that, you are commenting on your measurement method, not on the market.

Question 3: What do buyers complain about?

This is the most profitable of the three questions, and the answer is free: your competitors' customer reviews are a study somebody else already paid for.

Method: take the market's best-selling products and read the 2 and 3-star reviews. The 5-star ones teach you nothing. The 1-star ones are mostly about delivery and after-sales, rarely about the product.

Read until the same complaints recur. From that point you have your selling points: phrased by the market itself, which beats anything you would have invented.

What you do with it

The three answers combine into a single positioning decision:

• Concentrated market + empty price band + recurring complaint: the best configuration. You enter through the open space, with the flaw fixed.
• Fragmented market + tight prices + no clear complaint: the worst. You would be one more seller doing the same thing.
• Concentrated market + no gap + complaints about service: workable, but your difference will be operational rather than product-based. Know that before you launch, not after.

Competition is only one of the five criteria that decide whether a niche is worth it. The other four (demand, trend, margin, seasonality) are covered in finding a profitable niche.

The part everyone forgets: doing it again

A competitive analysis is a photograph. It is accurate on the day you take it, then it goes stale: prices move, sellers arrive, others leave.

What actually helps you decide is not the photograph but the movement. A competitor cutting prices over two months tells you something today's price does not. Which is why a modest measurement, repeated every week under exactly the same rule, beats a very thorough analysis done once.