Miroslava runs a custom framing shop in Denver, and she has been in business fourteen years. When a national chain opened two miles away three years ago she held her breath, then watched the chain struggle to serve the local art community the way she did. She kept most of her clients. Last year something subtler started happening: a younger competitor, a solo framer who had been open eighteen months, started showing up in the same Google search results. Same neighborhood, lower prices, strong Instagram presence. Miroslava felt it in her new client numbers before she could articulate it. She spent one afternoon with an AI assistant doing something she had never done systematically, mapping out exactly what her competitor offered, at what price, how he was positioning it, and where his strategy left gaps she could exploit. By the end of that afternoon she had a clearer picture of her competitive position than she had held in years.

Why This Is Suddenly Worth Doing

Competitive analysis used to take a research team or a consultant. For a small business it often just did not happen, because it was too time consuming and too expensive to justify against everything else demanding attention that week. AI changes the math. The gathering still takes you an afternoon, but the part that used to defeat owners, reading everything you collected and turning it into a picture you can act on, now takes minutes instead of a full day. That shift is what makes a meaningful competitive analysis realistic for a shop with no analyst on staff.

It is worth being precise about the division of labor, because this is where owners get disappointed. You do the collecting. An AI assistant working from what you paste in does the synthesis: comparing, grouping, naming patterns, and pointing at gaps. It is not out on the web checking your competitors for you, and it does not know your street. Feed it good raw material and it earns its place immediately. Feed it nothing and it will happily produce plausible sounding competitor analysis that describes no real business.

The Four Questions That Matter

At the enterprise level, competitive analysis involves market research firms, surveys, and financial reports, none of which are available to a shop with fourteen years of history and no analyst. Scaled down for a small business, the exercise means answering four practical questions, and everything else you might collect is either evidence for one of them or a distraction from all four. Write them down before you gather anything, because the questions are what keep an afternoon of research from turning into a folder of screenshots you never open again.

1. Who else is competing for my customers?
2. What do they offer and at what price?
3. How are they positioning themselves, and what story are they telling?
4. Where are they weak, and can I exploit that?

Notice that only one of the four is about price. Owners tend to collapse the whole exercise into a price comparison, decide they are more expensive, and stop there. The other three questions are where the useful answers live, because they tell you what a competitor is promising, who is believing the promise, and where the promise is breaking down. Keep the four questions written at the top of your working document and check every piece of information you gather against them.

Building Your Competitor Set

Start with a list. For most small businesses the relevant competitors fall into three groups: direct local competitors in your same geographic area and category, online alternatives your customers might use instead of you, and adjacent businesses that solve the same underlying customer problem by a different route. That third group is the one owners forget, and it is often where the quiet erosion comes from.

For Miroslava the set was concrete: the direct local framers, the national chain nearby, online custom framing services such as Simply Framed, and the younger competitor whose name kept appearing next to hers in search. That is a workable competitive set. It is small enough to research properly in an afternoon and broad enough that it includes the two threats that do not look like her business at all, the national chain and the online service.

Build your own list the same way, and be honest about the third group. Ask where the money would go if your business closed tomorrow. Some of it would go to the obvious rival across town, but some would go to a national option, some to an online service, and some to a completely different way of solving the problem that your customer would consider good enough. Each of those destinations belongs on the list. A competitive set made only of businesses that look like yours will confirm what you already believe and tell you nothing you did not know before the afternoon started.

Gathering the Raw Material

Now gather raw information about each competitor. This takes two to three hours of straightforward research, and it is deliberately unglamorous work. Open one document per competitor and paste everything in without editing or organizing it. Organising as you go is the single most common way this exercise stalls, because you start forming conclusions from the first competitor and then read the rest looking for confirmation.

• Their website: What services do they list? What prices are visible? What is the tone and the positioning?
• Google reviews: What do customers praise? What do they complain about? How recent are the reviews and what is the average rating?
• Social media: What content are they producing? How are they engaging with customers? What is their aesthetic and their voice?
• Their listing information: Hours, location, and services listed on their Google Business Profile.

Copy all of this into the document as you find it. Do not summarize, and do not clean it up. The raw wording of a competitor's website is evidence about positioning in a way your paraphrase of it is not, and the actual sentences customers wrote in reviews carry information that a tidy summary throws away. One document per competitor, everything in, judgment later.

Using AI to Find the Pattern

This is where AI earns its place. Take the raw information you gathered and paste it into an AI assistant with a prompt that establishes who you are, what you have attached, and what decision you are trying to make. Something in this shape works well:

"I am a custom framing shop in Denver. I have gathered information about my four main competitors below. Please analyze their positioning, pricing, strengths, and weaknesses. Then tell me where each competitor is vulnerable and where there might be opportunities for my business to differentiate."

Paste the raw competitor information after your prompt. What comes back in sixty seconds would take you two hours to synthesize manually: a structured comparison, identified patterns in how competitors position themselves, and explicit gaps in what the market is offering. Read it as a first draft of an argument rather than a verdict. Anything it asserts about a specific competitor should be traceable back to something you pasted in, and if you cannot find the source sentence, treat the claim as unproven and check it yourself.

Miroslava's analysis surfaced something she had noticed but never articulated. Every competitor was competing on speed or on price. Nobody was explicitly competing on local artist relationships and curation knowledge. That was her actual differentiation, and she was not communicating it anywhere on her website. The analysis did not tell her anything a very patient reader of those four documents could not have found. It found it in an afternoon rather than never.

Pricing Intelligence

Understanding competitor pricing helps you avoid two mistakes that pull in opposite directions: pricing so high that price sensitive customers never give you a chance, and pricing so low that you leave margin on the table and quietly signal low quality. Both are expensive, and the second one is harder to see, because a business that is underpriced usually looks busy.

For visible pricing, meaning anything listed on a website or a price sheet, gather it directly. For opaque pricing, where competitors do not publish rates, you have two options. Mystery shopping means calling or requesting a quote as a prospective customer would. Review mining means reading customer reviews for the prices people mention in passing, which they do surprisingly often when they are either delighted or annoyed. Neither gives you a full price list, and both give you enough to place yourself.

Once you have pricing data, ask an AI assistant to help you interpret your position rather than to set a number: "Based on this pricing data, am I priced at a premium, at parity, or below the market? Given my quality and service level, where should I be priced?" The useful output is the argument, not the figure. If the reasoning says you are positioned as a premium provider while your prices sit at the bottom of the market, that mismatch is the finding, and what you do about it is your decision.

Review Mining at Scale

Your competitors' negative reviews are your strategic intelligence, and they are public. Copy the one star and two star reviews from Google and Yelp for your two or three closest competitors, paste them into an AI assistant, and ask a direct question: "What are the most common failure patterns in these reviews? What do customers wish these businesses did differently?"

What comes back is a list of unmet needs your competitors are not solving. For each one, ask yourself two follow up questions. Can I actually solve this for customers, reliably, not just on a good week? And can I communicate that I solve it, on my website or in my marketing, in language a customer would recognize as an answer to the thing that annoyed them? A gap you cannot credibly fill is interesting. A gap you can fill and can describe is a sales asset.

Read the positive reviews too, but read them for a different purpose. Negative reviews tell you where a competitor is failing; positive reviews tell you what customers in your market have decided to care about, in their own words. When several people praise a rival for the same thing, that is the standard you are being measured against whether or not you chose to compete on it. It is also the most usable source of plain language you have, because customers describe what matters to them far better than any business describes its own offer.

Turning Findings Into Promises

Miroslava found that her main competitor's negative reviews clustered around two themes: delayed timelines, and staff turnover that meant customers worked with a different person on each visit. She turned both into explicit marketing messages. "Your artwork returned in two weeks, guaranteed" answers the first. "You will work with Miroslava directly, every time" answers the second. Neither claim is about being better in the abstract. Each one names a specific disappointment the market has already experienced and says, plainly, that it does not happen here.

That is the whole point of the exercise. A competitive analysis that ends with a document is unfinished. It ends properly when something visible to customers has changed: a line on your home page, a sentence in your quote email, a service guarantee you are prepared to honour. Before you publish a promise like that, check that you can keep it during your busiest month, not your quietest one, because a guarantee you break in your peak season generates exactly the reviews you were mining your competitor for.

How Often to Do This

A comprehensive competitive review like this is worth doing once or twice a year. Doing it more often produces churn rather than insight, because positioning does not move fast enough to reward monthly re-examination. Between reviews, lighter monitoring is valuable monthly and takes very little time:

• Set a Google Alert for each competitor's name, so you are notified when they are mentioned in news or reviews.
• Check their Instagram or Facebook once a month, looking for promotions, new services, or shifts in messaging.
• Watch your own Google search results for your primary service terms. A new competitor appearing in the top five is a signal worth investigating.

AI does not automate this ongoing monitoring for you. You still have to do the gathering. What it compresses is the synthesis step, and that is the step that used to make the whole exercise collapse. Keep the monthly monitoring light enough that you actually do it, and save the deep analysis for the once or twice a year when you have an afternoon and a real decision to make.

Anti-Patterns

Treating the AI's output as verified fact about a real business. A synthesis is only as good as what you pasted in, and an assistant asked about a competitor it has no information on will produce fluent, specific, wrong detail. Every claim in the output should trace back to a sentence in your raw documents. If it does not, it is a hypothesis you have to check, not a finding.

Gathering forever and synthesizing never. Some owners collect competitor material for weeks, feel productive, and never sit down to draw a conclusion. The research phase has a defined end: two to three hours, one document per competitor, then stop and analyze. More raw material past that point rarely changes the answer.

Using the analysis to copy. The instinct after seeing a competitor's strengths is to match them, feature for feature. That leads you into their positioning, where they are already established and you are the imitation. The output you want is the gap nobody is filling, which in Miroslava's case was not faster or cheaper framing but curation knowledge and artist relationships.

Reacting to every competitor move. Monthly monitoring is there to catch structural shifts, not to trigger a response to each promotion a competitor runs. If you rewrite your pricing every time a rival posts a discount, you have outsourced your strategy to someone whose business you do not understand from the inside.

Doing it once and calling it done. The competitor who is invisible to you today is the one who has been open eighteen months and has not reached your search results yet. A file from two years ago describes a market that has moved. Put the next review in your calendar before you close the document.

Practice Prompts

Work through these with your own raw material, not with hypotheticals. Each one assumes you have already pasted in the documents you gathered.

• Synthesis: "I run [type of business] in [location]. Below is the information I gathered on my main competitors. Analyze their positioning, pricing, strengths, and weaknesses, then tell me where each is vulnerable and where I could differentiate."
• Positioning check: "Based on this pricing data, am I priced at a premium, at parity, or below the market? Given my quality and service level, where should I be priced, and what is the reasoning?"
• Review mining: "These are the one star and two star reviews for my two closest competitors. What are the most common failure patterns? What do customers wish these businesses did differently?"
• Message drafting: "Here are the three most common complaints about my competitors, and here is what my business actually does about each one. Draft three plain sentences I could put on my website that answer those complaints without naming anyone."

Reflection

• Write down, right now and without research, who you think your three closest competitors are. When you build the real competitor set, how many of the three survive, and who did you miss?
• Of the four questions, which one have you genuinely never answered about your market: who competes, what they charge, how they position, or where they are weak?
• What are you actually better at than everyone in your competitive set? Now find where that claim appears on your website. If it appears nowhere, why not?
• What complaint appears repeatedly in your competitors' negative reviews that would also be a fair complaint about you on a bad week?

Glossary

Competitive set: The specific list of businesses your customers actually consider instead of you, covering direct local rivals, online alternatives, and adjacent businesses solving the same problem differently.

Positioning: The story a business tells about why it is the right choice, expressed through its pricing, its language, and its visible priorities rather than through a formal statement.

Review mining: Reading customer reviews systematically for evidence, including prices customers mention in passing and repeated complaints that reveal unmet needs.

Mystery shopping: Requesting a quote or making an enquiry as a prospective customer would, used to establish pricing that a competitor does not publish.

Price parity: Sitting at roughly the same price level as the rest of your market, as distinct from a premium position above it or a value position below it.

Differentiation gap: A customer need that your competitive set is collectively not serving, which is the output a competitive analysis exists to produce.

Related Lessons

• Competitive Intelligence and Market Positioning extends this afternoon exercise into an ongoing intelligence practice.
• Competitor AI Adoption Analysis looks specifically at what your rivals are doing with AI, rather than at their offer and pricing.
• Voice of Customer Analysis with AI applies the review mining technique to your own customers instead of your competitors'.
• Market Intelligence and Trend Analysis widens the frame from named competitors to the direction the whole market is moving.
• AI-Powered Pricing Optimization picks up where the pricing intelligence section stops, turning position into an actual price decision.

Closing

Miroslava did not learn anything in that afternoon that was hidden. Her competitor's website was public, his prices were public, and the complaints about the shop down the road had been sitting in Google reviews for anyone to read. What she lacked was not access but the hours to read it all and the structure to make sense of it. That is exactly the constraint AI lifts. Block out one afternoon, gather honestly, ask for the pattern, and finish by changing one visible thing about how you present your business.

Key Takeaways

• Competitive analysis for small businesses answers four questions: who is competing, what they offer and at what price, how they position themselves, and where they are vulnerable.
• Gather raw data first, then use AI to synthesize. Paste competitor website copy, pricing, and reviews into an AI assistant and ask for pattern analysis. Sixty seconds of that work replaces two hours of manual synthesis.
• You do the gathering; the assistant does the reading. It is not out on the web checking competitors for you, and any claim it makes should trace back to something you pasted in.
• Your competitors' negative reviews are strategic intelligence. Repeated complaints identify unmet needs you can explicitly promise to solve, provided you can honour the promise in your busiest month.
• Price positioning matters as much as the price itself. The finding you are looking for is a mismatch between the quality you claim and the price you charge.
• Do a thorough review once or twice a year and monitor lightly every month. Alerts, a monthly social media check, and watching your own search results catch shifts early.
• The goal is not to copy competitors but to find where they are weak. Build your strategy around the gap nobody is filling, not around matching what a rival already does well.

Frequently Asked Questions

How many competitors should I analyze? Enough to cover the three groups: direct local rivals, online alternatives, and adjacent businesses. Miroslava's set had four members, which is enough to see a pattern and few enough to research properly in an afternoon. If your list runs long, rank by how often you actually lose customers to each one and research the top of that list first.

Can the AI assistant go and look up my competitors for me? Treat the answer as no and structure your work accordingly. The reliable pattern is that you gather the material and paste it in. An assistant asked about a specific local business it has no information on can produce confident, detailed, invented answers, and a competitive analysis built on that is worse than no analysis at all.

What if my competitors do not publish their prices? Use the two indirect routes. Mystery shopping means requesting a quote as a prospective customer would, and review mining means collecting the prices customers mention in their reviews. Neither produces a complete price list, but together they usually give you enough to know whether you sit above, at, or below the market.

I did the analysis and found no gap. What now? That usually means the gathering was too thin rather than that the market is perfectly served. Go back to the negative reviews, which are the richest source, and read the actual sentences rather than the ratings. If the market genuinely is well served on every axis you can find, that is itself a finding, and the honest response is a decision about whether to compete on execution or to change what you sell.