The short version: You can use free AI tools plus about an hour of your own time to work out exactly why a competitor is beating you, what content is driving their traffic, and which of their tactics you can copy this week. Most small business owners skip this because they think competitor research needs expensive software. It doesn’t. It needs a method, and I’ll give you mine.
Why “just check their Instagram” isn’t research
I get asked some version of this question at least once a fortnight: “How do I find out what my competitors are doing that’s working?” And the honest problem isn’t that people don’t look at competitors. They look constantly. They scroll a rival’s Instagram, feel a bit sick, screenshot a post, and go back to their own desk with no plan. That’s not research, that’s anxiety with extra steps.
Real competitor analysis answers three questions: what content or product is bringing them traffic, what’s converting that traffic into money, and why their audience trusts them enough to buy. Screenshots don’t answer any of those. Data does.
The method I use with clients
This is the exact process I walked through with a client called Claire, who runs a corporate gifting business in the Midlands, sending branded hampers and welcome boxes to companies for staff onboarding and client thank-yous. Her biggest competitor was a bigger, better-funded company that seemed to be everywhere. Here’s what we did, in one sitting, in under 90 minutes.
- Step 1, pull their traffic sources (10 minutes). I used the free version of SimilarWeb to see the competitor’s top traffic channels. Free plan only shows you a slice of the picture, but it’s enough. For Claire’s competitor, 61% of visits came from organic search and a chunk of that organic traffic was landing on one single page.
- Step 2, find the page pulling the weight (5 minutes). That one page wasn’t a product page. It was a gift budget calculator. Type in headcount and budget per head, get a suggested hamper. It was ranking for terms like “corporate gifting budget calculator” and “how much to spend on employee gifts,” phrases nobody thinks to target but plenty of HR managers type into Google in November.
- Step 3, feed the findings to AI and ask it to reverse-engineer the strategy (20 minutes). I pasted the competitor’s homepage copy, that calculator page copy, and their top five blog titles into ChatGPT and asked it to map out the buyer journey they’d built: what search intent each page targets, what emotional trigger the copy uses, and where the gaps are. This is the part people skip. Most business owners either eyeball the competitor’s site themselves, or they ask AI to “analyse my competitor” with no inputs, and get back generic advice like “post more video content” that could apply to literally any business on earth. The output is only as sharp as what you feed it.
- Step 4, check what the competitor’s audience talks about (15 minutes). Free tools like AnswerThePublic or the free tier of SparkToro (25 lookups a month) show what questions and accounts a given audience segment engages with. For Claire’s market, HR managers were asking about “return to office gifts” and following productivity and workplace culture accounts, not gifting accounts. That told us where to place ads and guest content, nowhere near where Claire had been spending her budget.
- Step 5, decide what to copy and what to skip (30 minutes). This is the bit AI can’t do for you. We didn’t copy the calculator idea outright. We built a simpler one, a “cost per employee vs retention impact” calculator, because Claire’s actual selling point was retention data, not just gift selection. If you want to see how a tool like this drives real engagement rather than sitting unused, I’ve written before about how interactive calculators convert visitors better than static content, and it’s exactly the mechanism we leaned on here.
Total cost: zero pounds in software, about 90 minutes of time, and one afternoon of building. Claire’s calculator page now brings in roughly 800 organic visits a month and has become her second-biggest source of qualified leads, behind referrals.
The uncomfortable bit nobody says out loud
Here’s the thing I won’t dress up. Every single one of your competitors can run this exact same process, with the same free tools, in the same 90 minutes. The AI part of this isn’t a secret weapon anymore, it’s a level playing field. When everyone can ask ChatGPT to “analyse this competitor’s homepage,” everyone gets structurally similar output back, and the businesses that only do the AI step end up with strategies that look suspiciously alike. I’ve seen three small business websites in the same niche launch near-identical “free guide” lead magnets within a month of each other, clearly built off the same AI prompts.
What separates the businesses that pull ahead isn’t the analysis, it’s the judgment on top of it. Claire didn’t copy the calculator, she rebuilt it around her own actual proof point, which the AI had no way of knowing mattered because it doesn’t know her client testimonials or her retention data. The AI can tell you what a competitor built and roughly why it might work. It cannot tell you what’s true and provable about your business that makes a similar move credible. That gap is where the real advantage sits now, and it’s where most people stop trying because it takes actual thinking rather than a prompt.
What to look for when you study strong brands
The same reverse-engineering approach works on brands well outside your own market too, and it’s often more useful because there’s no emotional flinch. I’ve pulled apart plenty of these myself. Hotjar’s marketing strategy is worth studying for how a B2B tool used free versions of its own product as the top-of-funnel content engine. Duolingo’s approach shows how consistency and a bit of unhinged personality on social media builds recall far cheaper than paid ads. Calm’s strategy is a masterclass in partnership marketing, borrowing audiences from bigger platforms instead of building from zero. And GoPro’s marketing is the clearest example of a brand that turned its customers into its entire content team.
None of those companies are your competitor, and that’s the point. Feed their public content into an AI tool, ask it what mechanism is doing the heavy lifting, then ask yourself what the small-business version of that mechanism looks like for you. GoPro’s mechanism is “customers create the content.” A local personal trainer’s version of that might be asking every client for a 15-second phone video after their tenth session. Same mechanism, completely different budget.
How to do this without it eating your week
Set a hard limit. I do this for one competitor and one brand outside my industry per month, no more. Any longer and you’re procrastinating dressed up as strategy. My actual working rhythm:
- Monday morning, 30 minutes: pull traffic and top pages for one competitor
- Same session, 20 minutes: run the AI teardown prompt on their homepage and top three pages
- Same session, 10 minutes: write down one thing to test, one thing to ignore
That’s it. One hour, once a month, produces more usable direction than an entire afternoon of doom-scrolling a rival’s Instagram Stories.
If you’re running this process and consistently landing on “we need a content plan” or “we need better ads” without ever getting to specifics, that’s usually a sign the gap isn’t research, it’s implementation, and that’s a different conversation entirely. Some business owners get to this point and realise they need a second pair of hands to build the thing the research points to, which is when working with an AI consultant for small business tends to save more time than another month of solo research.
What I’d tell you to skip entirely
Don’t bother with expensive enterprise competitor intelligence software when you’re under 20 employees. I’ve trialled three of them for clients over the past two years, and the free-tier combination of SimilarWeb, AnswerThePublic, and a well-fed ChatGPT prompt got us 90% of the useful insight for 0% of the cost. The paid tools add speed and automation, not better thinking. And thinking is the part that’s in short supply.
Also skip the trap of only studying competitors your own size. The most useful teardown I did all last year was on a productivity app I’d never heard of, using IFTTT’s marketing strategy as the source material, because their whole approach to making a technical product feel simple through plain-language use cases translated directly into how I now brief AI prompts for non-technical clients. The lesson came from a completely unrelated industry, which is usually where the best ones hide.
Frequently asked questions
Do I need paid tools to reverse-engineer a competitor’s strategy?
No. The free tiers of SimilarWeb, AnswerThePublic, and SparkToro combined with a detailed ChatGPT prompt cover the vast majority of what a small business needs. Paid tools add automation and history, not sharper insight.
How often should I analyse competitors?
Once a month for a direct competitor and once a month for a brand outside your industry is enough. More frequent than that becomes a distraction rather than a strategy.
Isn’t it risky to copy what a competitor is doing?
Copying the exact tactic rarely works because you’re missing their proof points and audience trust. Copying the underlying mechanism, then rebuilding it around your own strongest evidence, is what converts.
What should I feed the AI tool during a teardown?
Paste the competitor’s actual homepage copy, their top three ranking pages, and a handful of their blog titles. Vague requests like “analyse my competitor” with no source material produce generic advice that fits any business and helps none.
Related reading: Why Your AI Content Sounds Like Everyone Else’s (And It’s Not the Tool’s Fault) and Impossible Foods Marketing Strategy: How They Built a Brand That Wins.
It also helps to compare who AI answers cite for the same questions, so you can see where a competitor is winning visibility you are not.
For the bigger picture, see my full guide to AI marketing.
