How to track competitor keywords and see if they are beating you

Two different questions: what your competitors rank for vs who is beating you on your terms

When someone sets out to track competitor keywords, they usually have one of two needs in mind. The two look similar on the surface and share nothing underneath. Confusing them is expensive: you buy the wrong tool, run the wrong analysis, and end up with conclusions that do not answer your original question.

The first question is discovery: which queries does my competitor rank for that I am not even working on. That is a market question. You answer it with a keyword database that has crawled millions of queries and can tell you, with a meaningful margin of error, which terms drive traffic to somebody else’s domain. It is research: it gives you content ideas, market gaps, topics you had not considered.

The second question is surveillance: on the queries I already know matter, who is above me and since when. That is an operational question. You do not need millions of queries; you need twenty, fifty or a hundred specific keywords, measured every day, with memory attached. This is the question a rank tracker answers, and it is the one this guide develops.

Why insist so hard on the distinction? Because the SEO competitor analysis that actually moves the needle for a small or mid-sized site almost never starts with discovery. It starts with the ground you already work. If you sell invoicing software and twelve queries drive 80% of your signups, learning that a competitor ranks for something adjacent is interesting, but learning that last week they pushed you from position 4 to 7 on your main query is actionable today.

There is also a data quality difference. Keyword databases work in estimates: estimated traffic, estimated position, estimated volume. Useful for orientation, unreliable for any decision that hinges on an exact number. Top 10 surveillance is a direct measurement: query the SERP, read the result, store it. No statistical model in between. When the report says domain X sits at position 3, it is because it sat at position 3 on the day and in the country you measured.

Before going further, the full map of what can be measured continuously lives on the Rankiamos features page. And one clarification that saves confusion later: if you find that two of your own URLs are getting in each other’s way on the same query, that is not a competitor problem but an internal one, with a different method — covered in the guide on detecting keyword cannibalization with Search Console.

What each approach answers, and how precisely

This is the table I pull up whenever somebody asks for "a tool to spy on competitor keywords." Before recommending anything, you have to know which column solves the real problem.

Nobody forces you to pick one. The sensible setup on a serious project is a keyword database a couple of times a year, when you plan content, and a tracker all year round, because the ground shifts weekly. What makes no sense is paying for an expensive research subscription and using it as a tracker, or expecting a tracker to do market research.

Why today’s top 10 tells you nothing without the last few weeks

Open Google, search your main query, write down the ten results. You now have the same data any tool on the market would give you today. And yet that data, on its own, does not support a single decision. An isolated top 10 has no direction: you cannot tell whether the domain at position 4 has been there for two years or arrived the day before yesterday, and those two situations demand opposite responses.

The top 10 of a normal commercial query moves constantly. Google tests variations, fresh results rotate in, pages that lost relevance drop out, SERP features toggle on and off and push organic results down. A large share of that motion is noise: one-position changes that revert on their own within three days. Without history, every look at the SERP is a random sample you are mistaking for a state.

Take a concrete example. Your query is "invoicing software" and today the top 5 looks like this: your site at 4, a large competitor at 1, an accounting blog at 2, a domain you have never heard of at 3, and a marketplace at 5. What do you do? There is no possible answer from that alone. Now add eight weeks of history:

All three scenarios produce exactly the same snapshot today and demand three different decisions. That is what it means to say history is not an extra: history is the data. The snapshot is the byproduct.

There is an extra wrinkle that hits multi-market sites hard: results change by country. You can hold position 3 in one market and 11 in another for the same query, because Google adapts the SERP locally. In our own Search Console data over the last 90 days, our average Google position differs sharply between countries for identical queries — Colombia averages considerably better placements than the United States. If you run competitor analysis off a single national result while selling in three countries, you are watching ground that is not yours.

And here comes the honest constraint you have to plan around: your own history starts the day you add the keyword. It is not retroactive and there is no way to rebuild it. If you are shipping new content in September and want to prove in November what happened to the top 10 of that query, the keyword needs to be tracked from August. This is the most frequent planning mistake I see, and it cannot be fixed after the fact.

A competitor ranking history is, at bottom, a table of domains by date with the position they held in the top 10 of one of your queries. That sounds simple, and it is. What is not simple is reading it without over-interpreting. These are the five patterns worth distinguishing, ordered from most to least urgent.

1. Pattern 1 · Sustained climb

   A domain improves in consecutive steps over three weeks or more, never falling back by more than one spot. This is the strongest signal on the board: Google keeps finding more reasons to prefer it. There is almost always an identifiable cause behind it — new content, a page merge, a performance fix, recent links. It is the only pattern that justifies changing your quarterly plan.

2. Pattern 2 · New entry that sticks

   A domain that was absent enters the top 10 and is still there two weeks later. Persistence is the key, not the entry: entries happen constantly and most last days. If it is still inside after fifteen days, it is no longer a Google experiment, it is a competitor. Note it and check which specific URL is ranking, because it is often an article rather than the homepage.

3. Pattern 3 · Oscillation without trend

   A domain bounces between 5 and 9 with no clear direction for weeks. That is noise and the correct action is none. Most day-to-day SERP movement falls in this bucket. React to every oscillation and you will be rewriting content forever with no measurable result.

4. Pattern 4 · A rival dropping steadily

   Someone who held the top 3 for months starts sliding consistently. That is an opportunity and it usually goes unnoticed, because nobody watches competitors while they are doing well. A gap opening at position 2 or 3 on a commercial query is more profitable than attacking a brand new query from zero. If you sit at 6 or 7 and the one above you is falling, that is the moment to strengthen that page.

5. Pattern 5 · Whole top 10 rotating

   It is not one domain moving: it is nearly all of them in the same week. That is rarely anyone’s merit or fault in particular; it usually points to an algorithm update or a change in SERP composition. The correct reading is contextual: if everyone moved and you moved too, your drop is not yours. It is the check most people skip before panicking.

A worked reading, with numbers

Query: "seo agency bogota." Eight weeks of daily data, five domains on the board. Your site starts at position 5 and ends at 6. Framed that way, nothing important seems to have happened. Look closer: domain A went from 8 to 3 in one- and two-step moves, never falling back more than one spot; domain B bounced between 2 and 4 every week; domain C vanished from the top 10 in week 5 and never returned; domain D entered in week 6 at position 9 and is still there.

The conclusion is not "I lost one spot." The conclusion is that domain A is a sustained climb worth studying today, that B is noise and deserves no attention, that C leaving freed a slot which A — not you — absorbed, and that D is not yet anything. Your slide from 5 to 6 was not an attack: it was the side effect of someone else rising. That is a completely different conclusion and it leads to a different plan.

Notice something important about that example: the useful work was not looking at your own line. It was looking at everyone else’s. Most ranking reports show only your own curve, and your own curve never explains why it moved. The "why" always lives in the rest of the top 10.

One special case deserves its own treatment: when your line hovers between positions 11 and 20 for months while the top 10 rotates above you. There the problem is not a specific competitor but the entry threshold, and the levers are different — worked out in the guide on moving from positions 11-20 to page one.

The competitor showing up in 3 of your keywords that you had never heard of

This is, to me, the most valuable finding continuous top 10 monitoring produces, and the one no manual check ever surfaces. The question is not "who is above me on this query." The question is "who is outranking me on Google across several queries at once."

When you review keyword by keyword, each SERP is its own island and unfamiliar domains pass as background noise. When you cross twenty or fifty queries, a different pattern emerges: some domains appear in exactly one of your queries by coincidence, and others appear in three, five or eight. Those second ones are your real search competitors, even if they do not compete with you in the market.

And here is the uncomfortable part: they are almost never the companies you considered your competition. In practice, the domains that show up most tend to fall into these buckets.

The practical reading is straightforward. List every domain that appeared in the top 10 of your queries over the last month and count how many queries each one shows up in. Sort descending. The top three are your real search competition, and they earn an hour of your time: what content format they use, how deep their pages go, how often they update, whether they attack the query with a dedicated page or with the homepage.

One detail that changes the analysis a lot: pay attention to which URL is ranking, not just the domain. A competitor beating you with a 3,000-word guide written two years ago and one beating you with a landing page shipped last month are opposite diagnoses. In the first case Google is rewarding depth and age; in the second, freshness and commercial fit. Your response should differ.

And since we are being honest about limits: what this analysis does not give you is the link profile of those domains. Rankiamos does not measure backlinks, yours or anyone else’s. If your diagnosis requires knowing how many links a competitor earned last month, you will need another tool. What you do get is the observable effect of that authority, which is sustained position over time, and in most content decisions that is what ends up carrying the weight.

What you can see for free and where it starts costing money

A good chunk of basic SEO competitor analysis costs nothing. It is worth knowing exactly where free stops before deciding whether you need more, because a lot of content out there exaggerates the need for paid tools in order to sell subscriptions.

What you get for free

An incognito Google search gives you today’s top 10. That is real data and it costs nothing. The site: operator shows which pages of a domain are eligible for a topic. Google Search Console, which is free, tells you which queries you appear for, with how many impressions and at what average position — it says nothing about competitors, but it defines your ground, which is the starting point for everything else. And a spreadsheet with date, query and ten domains is a perfectly valid manual tracker if you have three queries.

The Rankiamos Free plan belongs in this category: 2 projects, 5 keywords, weekly automatic checks, the top 10 of each keyword with 45 days of history, Google Search Console connection and site scanning, no card required. To watch the single most important query of a small business and see who occupies the top 10 week over week, it is enough.

Where it starts costing

The free method breaks at three specific points, and all of them are about scale and time rather than the quality of any single measurement.

1. 1 · Past five queries

   Checking five SERPs by hand once a week takes fifteen minutes. Checking forty takes two hours, which is why nobody sustains it beyond a month. The real breaking point is not difficulty, it is consistency: the value of history comes from not missing a single week.

2. 2 · When you need daily resolution

   Four measurements a month show a coarse trend but cannot date a change. If you shipped content on the 12th and a competitor moved on the 14th, weekly data will never let you connect the two. The Pro plan’s daily checks are what provide that resolution.

3. 3 · When you measure several countries or cities

   Here manual work simply stops being reliable: your browser will not show you the Mexico City SERP from Bogota without tricks, and the tricks return inconsistent results. Mind the quota: each keyword-by-city combination consumes 1 unit. Adding cities does not raise the price, but it does consume quota.

The Pro plan costs $14.99 per month or $119.99 per year — $9.99 monthly equivalent — and includes 5 projects, 50 keywords and daily checks. The features that make continuous competitor analysis possible run on it: competitor movements, search volume, AI analysis (30 a month), citations in Google’s AI Overview, your position in the map pack of Google results and cannibalization detection. Competitor history comes with both plans. There is a 7-day free trial through Apple; it needs a payment method on your Apple account and renews unless you cancel before it ends. Full limits are on the pricing page.

And the platform limitation, stated before rather than after: Rankiamos runs on iPhone with iOS 17 or later, iPad, Mac with macOS 14 or later, and Apple Watch. There is no web version and no Android app, and the app's interface is in Spanish only. If your team works on Windows or a client needs to log into their own dashboard, that is not covered today. What you can do, on Pro, is export a PDF with your name and your color — no custom logo, to be clear — or an unbranded CSV, and send it yourself.

Having history in front of you creates a strong temptation: do something every time the line dips. That impulse is costly. Each response burns hours of content or engineering time, and most SERP movements revert on their own whether you act or not. The discipline is having a written threshold before you open the report.

This is the decision framework I use. It is not universal law, it is a reasonable starting point you can tune to how volatile your sector is.

Three questions before you lift a finger

Before approving any response to a competitor movement, answer these three. If any of them fails, there is no action.

First: has the movement lasted at least three weeks? Below that you do not have evidence of a trend, you have a sample. The one reasonable exception is a sharp, large drop of your own, which is worth investigating immediately because it usually has a technical cause.

Second: is the query worth the effort? Not every keyword in the report deserves defending. A low-volume query with no commercial intent can be lost without consequence. Before committing two weeks of content work, check what that query actually represents in your business. Rankiamos’ search volume and estimated traffic value, on Pro, help size it, but the business judgment is yours.

Third: do you have a concrete hypothesis for why they beat you? "I will improve the content" is not a hypothesis. "Their page covers the five questions mine answers in two lines" is one. Without a specific hypothesis, the response becomes rewriting for the sake of rewriting, and that rarely moves a position.

When the problem is not the competition

A sizable share of drops blamed on competitors are self-inflicted. Rule those out before writing a single new line. The three most common: two of your own URLs fighting over the same query, a recent technical change that broke something, and confusing the aggregated average position with the real SERP position.

The first has its own method, in the guide on detecting cannibalization with Search Console. The third is subtler than it looks: a worsening average position does not always mean you fell — sometimes it means you started appearing for new, worse-placed queries. That distinction is unpacked in average position vs real position. If your competitor diagnosis starts from a misread number, everything you build on top of it will be crooked.