Competitor analysis is the work of regularly recording what your competitors do and tying that record to your own decisions. It is not a one-off deck. A competitor analysis done once and filed away becomes a source of wrong information within three months.
What follows is a setup that works regardless of team size. Six steps, in order.
What competitor analysis is not
Competitor analysis is not imitation. The point is not to do what a rival does, it is to work out whether their move changes your plan. A team that answers every move has handed its roadmap to someone else.
It is also not a one-off deck. The value shows up in the series over time, not in a single snapshot. A competitor's price today is information. Six months of their price movement is intelligence.
1. Narrow the competitor list
Most teams start wide and drown in twenty competitors. It works better to split them into three groups. Direct competitors sell the same product to the same customer. Indirect competitors serve the same need with a different product. Aspirational competitors are bigger than you and show where the market is heading.
Watch the direct ones closely, review the indirect ones monthly, and follow the aspirational ones only on product and pricing moves. The tighter the list, the longer the habit survives.
Revisit the list every six months. A new entrant joins it, a rival that no longer sells to your customer leaves it. A list that never gets updated is the quietest reason teams end up watching the wrong companies.
2. Write the question before you gather anything
Starting with data collection is the most common mistake. Write the question first: where does our price sit in the market, is a competitor entering a new segment, is our campaign cadence enough. Without a question, the data you gather will not be used.
A good question is one whose answer can be attached to a decision. 'What is the competitor doing' is a bad question. 'How many times has the competitor changed its welcome offer in the past three months' is a good one.
3. Split your sources into three buckets
Competitor data arrives from three places, and they should not be mixed.
- Public surfaces: the website, pricing page, campaign pages, app store listings, job postings, press releases.
- Marketing channels: ad libraries, push notifications, email and SMS flows, social announcements.
- Indirect signals: customer reviews, support forums, the objections you hear in sales calls.
4. Define fields you can measure
The first bucket is the most reliable because it can be verified. The third is the richest but cannot carry a decision on its own. Whichever bucket it comes from, record every competitor against the same fields: price range, campaign type, offer size, target segment, channel, date.
Competitor notes written as free text become documents nobody opens six weeks later. A record split into fields becomes a table, and the table becomes a trend line.
Decide how many fields you will keep, too. A forty-field template gets filled in week one and sits empty by week three. Eight to twelve fields is the sustainable limit for most teams.
5. Set frequency and ownership
Every field needs a scan frequency and an owner. A pricing page can be weekly, campaign pages daily, job postings monthly. If the owner is not written down, in practice nobody watches.
When choosing frequency, ask what you would lose by learning this a day late. If the answer is nothing, lower the frequency.
Do not pile ownership onto one person. The pricing field belongs to someone in pricing, the campaign field to someone in marketing. People keep watching data that serves their own work.
6. Close every finding with a decision
The output of competitor analysis is a decision, not a report. Every finding should end in one of three states: we are acting, we are watching, we are closing it as unimportant. The third is also a decision and it needs to be written down.
A finding without a decision gets re-litigated at the next meeting. That is the real reason teams get tired of competitor analysis.
Write the date and the reasoning next to the decision. When the same move reappears six months later, you will not have to remember why you sat it out.
What to build in the first 30 days
Do not try to stand this up in one go. Thirty days is enough of a start for most teams.
- Week 1: cut the direct competitor list to three and write the two questions you want answered.
- Week 2: define the fields, build the empty table and fill the first record by hand.
- Week 3: set the scan frequency and the owner for each public surface.
- Week 4: close the first findings with decisions and mark the fields that never moved. Drop those fields.
Three common mistakes
The third mistake deserves a post of its own: Why isn't 'the page changed' enough?
- Recording only the headline: the percentage gets noted, the conditions do not, and the comparison comes out wrong.
- Keeping no evidence: once the page changes, the detection can no longer be verified.
- Treating every change as equally important: noise rises and the team stops reading alerts.
With Adversee
Adversee automates the public-surfaces part of this setup: it scans competitor sites on a schedule, splits campaigns into fields, scores aggressiveness and keeps the evidence. For the hands-on version, step by step: How to track competitor campaigns?
Start tracking competitor moves automatically.