How to Use Claude for Competitor Analysis

Most solo operators doing competitor research waste 3–4 hours copying and pasting data from websites, social profiles, and listing portals into a spreadsheet that’s already outdated by the time they finish. I know because I did exactly that for the first decade of my real estate consulting business in Madeira. Then I started using Claude for competitor analysis, and that 4-hour grind dropped to under 45 minutes. Here’s exactly how I do it — step by step, no theory.

This guide walks through a repeatable process I’ve tested across real property market research. The steps work whether you’re analyzing competing real estate agencies, rival consultants, or any service business where you need to understand positioning, pricing, and messaging fast.

What Claude Can Actually Do for Competitor Analysis

Before jumping into steps, let’s be clear about what Claude is doing here. Claude doesn’t scrape the web in real time — it’s not Perplexity AI. What it does exceptionally well is structure, synthesize, and extract patterns from text you feed it. That distinction matters a lot for how you run this process.

You supply the raw material — competitor website copy, About pages, listing descriptions, social media posts, Google review text, pricing pages. Claude then does the analytical heavy lifting: identifying positioning angles, spotting messaging gaps, extracting implicit pricing signals, and building comparison frameworks.

For competitor analysis specifically, I use Claude Pro ($20/month) because the extended context window lets me paste in large volumes of competitor text in one session without constantly starting over. The free tier works for smaller comparisons but you’ll hit limits fast.

Step 1: Define Your Competitor Set Before You Open Claude

This sounds obvious. Most people skip it and regret it.

Before you touch Claude, write down exactly who you’re analyzing and why. I keep a simple working document for each competitor sprint I run — usually every quarter. The document lists:

• Direct competitors: businesses doing the same work, targeting the same clients, in the same geography
• Indirect competitors: businesses solving the same problem differently (for me, that’s online real estate platforms that bypass consultants entirely)
• Aspirational competitors: businesses one tier above yours that you want to understand

For Madeira specifically, my direct competitor list usually runs 6–8 agencies and independent consultants. I pick the top 4–5 for any given analysis session. Going broader than that dilutes the output — Claude still performs, but the insights get generic fast.

Write your competitor list. Include their website URLs. That’s all you need before Step 2.

Step 2: Gather Raw Competitor Data the Right Way

Open each competitor’s website and manually copy the text from these pages into a single document:

• Homepage (especially hero section and value proposition)
• About/Team page
• Services or listings page
• Any visible pricing page
• 3–5 recent social media posts if they’re active
• Their Google Business profile description and a sample of recent reviews

Yes, this is manual. That’s intentional. The data collection phase takes me about 20–25 minutes for 4 competitors. I label each section clearly: “COMPETITOR A — Homepage Copy” and so on. Clean labeling is what makes Claude’s output actually useful rather than muddled.

One thing I do not do: paste in URLs and ask Claude to “check their website.” It can’t browse live sites reliably enough for analysis I’d act on. Feed it the text directly.

Step 3: Build Your Analysis Prompt with a Clear Framework

This is where most people get mediocre results — they paste everything in and ask “what do you think?” Claude needs structure to return structured analysis.

Here’s the prompt template I use, adapted from what I’ve refined over about 14 months of regular use:

Below is website and social media copy from [NUMBER] competitors in [YOUR MARKET/NICHE]. 

Analyze each competitor across these dimensions:
1. Core value proposition (what problem they claim to solve, for whom)
2. Tone and messaging style (formal, casual, expert-led, lifestyle-focused, etc.)
3. Implied pricing position (premium, mid-market, budget — based on language signals)
4. Differentiation claims (what they say makes them different)
5. Gaps or weaknesses visible in their messaging

After analyzing each individually, give me:
- A side-by-side comparison summary
- The 3 biggest messaging gaps across all competitors combined (opportunities I could fill)
- 2–3 positioning angles I could adopt that none of them are using

Here is the competitor data:

[PASTE YOUR LABELED TEXT HERE]

The numbered dimensions in the prompt are non-negotiable for me. They force Claude to give me comparable output across competitors, which makes the later synthesis actually actionable.

Step 4: Run the Analysis and Interrogate the Output

Paste your prompt plus the competitor data and run it. Claude will return an initial analysis — usually well-structured if your prompt was clean. But don’t stop there.

The first output is the starting point, not the deliverable. I always follow up with at least two interrogation prompts in the same conversation:

Follow-up 1 — Challenge the gaps: “For each messaging gap you identified, give me a concrete example of the exact language or content type that would fill it — as if you were writing it for my business.”

Follow-up 2 — Find the blind spots: “What are competitors doing well that I should be aware of, even if I don’t copy it? What would a potential client see in Competitor B that might make them choose them over me?”

That second follow-up is uncomfortable. It’s also where I’ve gotten some of my most useful information. One time it told me flat out that a competing agency’s review volume on Google was so much higher than my own that any client doing basic due diligence would notice immediately. That was accurate. I spent the next month fixing it.

Step 5: Build a Competitor Comparison Table Inside Claude

Once the analysis is done, I ask Claude to format the findings as a comparison table I can drop straight into Notion or a client report. The prompt is simple:

“Now take everything from your analysis and build an HTML comparison table. Rows = the 5 dimensions we analyzed. Columns = each competitor plus a column for my business (leave those cells blank for now). Keep each cell under 20 words.”

Here’s what a typical output looks like — I’ve anonymized the competitor names:

Dimension	Agency A	Agency B	Consultant C	My Business
Core Value Prop	Volume of listings, fast transactions	Luxury focus, international clientele	Local knowledge, personal service	[Fill in]
Tone	Corporate, transactional	Aspirational, lifestyle-driven	Warm, conversational	[Fill in]
Pricing Signal	Mid-market, fee not mentioned	Premium, price implied by imagery	Budget-friendly signals, emphasizes value	[Fill in]
Differentiation Claim	“Largest portfolio in the region”	“Exclusive properties, curated selection”	“Born and raised here”	[Fill in]
Visible Gaps	No educational content, no trust signals	Cold tone, hard to contact	Thin online presence	[Fill in]

That table takes Claude about 30 seconds to produce. Building it manually from research notes used to take me the better part of an afternoon.

Step 6: Translate Findings into Actionable Strategy

Analysis without action is just expensive procrastination. Final step: ask Claude to convert the gap analysis into a concrete to-do list.

My prompt: “Based on the 3 messaging gaps you identified, give me a prioritized list of 5 specific actions I could take in the next 30 days — website copy changes, content pieces to create, or positioning language to test. Be specific enough that I can assign each one to a day’s work.”

The output from this step is where competitor analysis actually earns its time. Without this translation layer, most competitor research sits in a document and influences nothing.

My Real-World Experience Running This in Madeira’s Property Market

In January 2026 I ran a full competitor analysis sprint ahead of updating my service positioning for the year. I had 5 competitors I wanted to examine: three established agencies with physical offices in Funchal, one remote consultant who’d been gaining visibility through Instagram, and one large Portuguese platform that had expanded into Madeira’s residential market.

I collected text manually from all five — homepages, About pages, services listings, and Instagram bios where relevant. Total collection time: 28 minutes. I ran the full process through Claude Pro, starting with the structured analysis prompt, then two follow-up interrogation prompts, then the table, then the action list.

Total time from opening the blank document to having a formatted comparison table and a 5-item action list in Notion: 52 minutes. The same exercise the previous January — done manually with colored spreadsheet tabs and my own notes — took me nearly a full working day, around 6.5 hours.

The most valuable insight from the Claude analysis wasn’t something I would have spotted quickly on my own. It flagged that four out of five competitors were using almost identical language around “local expertise” — which meant that phrase had become invisible noise to anyone reading real estate consultant websites in Madeira. Nobody was talking about the process of working with a consultant: what happens at each stage, what to expect, how decisions get made. That was the gap.

I rewrote my services page in February to lead with a 4-step process section explaining exactly what clients experience when they work with me. Traffic to that page increased, and I’ve had three prospects in 2026 specifically mention they chose to contact me because “your website was the clearest about how it actually works.” That came directly from the Claude analysis flagging a gap none of my competitors had filled.

I want to be honest about what Claude did NOT do well in this process. When I asked it to estimate competitor fee structures based on language signals, the output was too speculative to use. It flagged one competitor as “budget-positioned” based on phrases like “flexible solutions” — but when I spoke to a colleague who’d worked with that agency, their fees were actually above market rate. Language signals alone are unreliable for pricing inference, and Claude was overconfident in that section. I now skip the pricing inference step unless I have actual pricing data to feed it.

Pro Tips From 14 Months of Running This Process

Run analysis quarterly, not once

Competitor positioning shifts. The Instagram-focused consultant I analyzed in January had almost no blog content. By April she had 8 posts. A one-time analysis gives you a snapshot, not a strategy. I calendar a competitor sprint every 90 days — it takes under an hour now that I have the process dialed in.

Add client review text for a second layer of insight

Google reviews from competitor profiles are public. Paste in 10–15 reviews from a competitor alongside their website copy and ask Claude: “What do customers actually value about this business — and does it match what the business claims to offer?” The gap between what a competitor says and what their clients say is often more revealing than the copy itself.

Save your prompts as a reusable template

I keep my full prompt sequence in a Notion page labeled “Claude Competitor Analysis Template.” Each quarter I open it, update the competitor data, and run the same sequence. No rebuilding from scratch. This is probably where I save the most cumulative time — the first sprint took me about 90 minutes including prompt refinement. Now it’s consistently under an hour.

Don’t ask Claude to evaluate your own business in the same prompt

I tried this early on — including my own website copy alongside competitors and asking for a full comparison. The output was noticeably more diplomatic about my own positioning than about competitors. Keep your self-assessment as a separate conversation or a separate prompt where you explicitly ask for harsh critique.

What Claude Tools Work Best for This Process

For competitor analysis specifically, I work almost entirely in the standard Claude chat interface on the Pro plan. I’ve experimented with Claude’s Projects feature (available on Pro) to store recurring competitor notes across sessions — that’s genuinely useful if you’re tracking the same competitors over multiple quarters and want Claude to remember prior context.

I don’t use Claude’s built-in tools (like web search when it’s available) for this workflow. The process works better when I control the data input — I know exactly what text Claude is working with, which makes the output more reliable and easier to verify.

Practical Summary and Next Steps

Here’s the full process compressed:

1. Define your competitor set before opening Claude (3–5 competitors maximum per session)
2. Manually collect text from key pages — homepage, About, services, reviews
3. Use a structured prompt with defined dimensions — don’t ask open-ended questions
4. Interrogate the initial output with follow-up prompts that challenge and pressure-test
5. Ask for an HTML comparison table for easy documentation
6. Translate findings into a 30-day action list before closing the session

The pricing inference limitation is real — treat any Claude output about competitor fees as speculative unless you’re feeding it actual pricing data. Everything else in this process has held up well across 6 quarterly analysis sprints in my own business.

If you’re running a solo consulting or service business and still doing competitor research manually, this workflow will cut that time by at least 70%. That’s not a guess — it’s what I’ve measured across my own quarterly sprints in Madeira’s real estate market.

Try the Step 3 prompt template on just one competitor this week. Run it, see what Claude surfaces, then come back and add the follow-up interrogation prompts. The process builds on itself — and the first session will show you immediately whether this is worth the 52 minutes I’m claiming it takes.