# Competitive Intelligence with Fish.Dog + Claude Code

Canonical HTML: https://fish.dog/claude-code-guide/competitive-intelligence-guide
Markdown mirror: https://fish.dog/claude-code-guide/competitive-intelligence-guide.md

LLM summary: Build competitive battlecards using a 7-question perception study that captures awareness, consideration, strengths, weaknesses, and switching triggers in 45 minutes.

Primary claim: Complete Claude Code guide to building competitive battlecards using Fish.Dog synthetic research. 7-question study design, full API workflow, multi-competitor parallel studies, quarterly tracking, and battlecard generation.

**New to this workflow?** Read the companion article first: [How to Build Competitive Battlecards with Claude Code and Fish.Dog](https://fish.dog/news/how-to-build-competitive-battlecards-with-claude-code-and-ditto). It explains the concepts in plain language before you dive into the technical implementation below.

A complete Claude Code guide to building competitive battlecards from synthetic perception research. Seven-question study design, full API workflow, multi-competitor parallel studies, quarterly tracking, and automated battlecard generation. From zero to finished battlecard in 45 minutes.

**For Claude Code agents and AI coding assistants.** This guide provides copy-paste API calls, study designs, and proven workflows for competitive intelligence research with Fish.Dog's synthetic research API. Every command is production-tested. The output is a finished competitive battlecard ready for sales distribution.

Contents

1. [The Three Layers of Competitive Intelligence](#three-layers)
2. [Anatomy of a Competitive Battlecard](#battlecard-anatomy)
3. [The 7-Question Competitive Perception Study](#7-question-study)
4. [Complete API Workflow: Step by Step](#complete-api-workflow)
5. [Generating the Battlecard from Study Data](#generating-the-battlecard)
6. [Multi-Competitor Parallel Studies](#multi-competitor)
7. [Quarterly Competitive Tracking](#quarterly-tracking)
8. [Advanced: Claim Credibility Testing](#claim-credibility-testing)
9. [Worked Example: SaaS Analytics Tool vs Mixpanel](#worked-example)
10. [Connecting CI to Other PMM Workflows](#connecting-to-other-workflows)
11. [Best Practices and Common Mistakes](#best-practices)
12. [Frequently Asked Questions](#faq)

---

## 1. The Three Layers of Competitive Intelligence

Competitive intelligence operates across three layers. Fish.Dog addresses the third, which is the hardest to obtain through existing tools.

| Layer | What It Tells You | Existing Tools | Gap |
| --- | --- | --- | --- |
| **Layer 1: Data Collection** | What the competitor is *doing*: website changes, pricing updates, job postings, press releases, feature launches | Klue, Crayon, G2, Owler, SimilarWeb | Well served |
| **Layer 2: Business Context** | What the competitor is *planning*: hiring patterns signal strategy shifts, pricing changes signal repositioning | Klue, Crayon, analyst reports | Adequately served |
| **Layer 3: Customer Perception** | What the market *thinks* about the competitor: do buyers believe their claims? What triggers switching? How do they compare you? | Win/loss interviews (slow, small sample, hard to recruit) | **This is where Fish.Dog fills the gap** |

**Why Layer 3 matters most:** 68% of sales opportunities are competitive. In those conversations, the outcome depends on how well your team understands buyer perception of the competitor. Layer 1 and 2 data tells you what the competitor claims. Layer 3 tells you whether buyers believe it.

---

## 2. Anatomy of a Competitive Battlecard

The study design in this guide produces all six sections of a standard competitive battlecard. Each section maps to specific study questions.

| Section | What It Contains | Source Questions | Who Uses It |
| --- | --- | --- | --- |
| **Why We Win** | Top 3-4 reasons customers choose you over this competitor, in customer language, with evidence | Q2, Q5, Q7 | AEs in competitive deals |
| **Competitor Strengths** | Honest assessment of where the competitor excels + how to respond when prospects raise these points | Q3 | AEs, SEs |
| **Landmine Questions** | High-impact questions that expose competitor gaps without sounding adversarial ("Have you asked them about X?") | Q3 (weaknesses), Q4 (scepticism) | AEs in discovery calls |
| **Quick Dismisses** | 1-2 sentence rebuttals for the competitor's most common claims | Q4 | AEs, SDRs |
| **Switching Triggers** | Events and frustrations that create openings: price increases, buggy releases, lost integrations | Q6 | AEs, SDRs, Marketing |
| **Recent Wins** | Specific deals won against this competitor with industry, use case, and deciding factor | Supplemented from CRM data | AEs for social proof |

### The ABC Quality Framework

- **Accuracy:** One inaccuracy destroys seller trust and the card gets binned. Every claim must be evidence-backed.
- **Brevity:** One page maximum. Scannable in 60 seconds during a live call. A 10-page competitor dossier is less useful because no one reads it.
- **Consistency:** Standardised format across all competitor cards. Reps should know exactly where to find each section.

---

## 3. The 7-Question Competitive Perception Study

Each question targets a specific competitive intelligence need. Together, they produce raw material for all six battlecard sections.

| Q# | Question | CI Component | Battlecard Section(s) |
| --- | --- | --- | --- |
| 1 | "When you think about solutions in `[category]`, which brands or tools come to mind first? What do you associate with each?" | Brand Awareness + Associations | Competitive landscape context |
| 2 | "You are evaluating `[your product]` against `[Competitor A]`. What would make you lean toward one or the other?" | Head-to-Head Decision Drivers | **Why We Win** + Competitor Strengths |
| 3 | "What is the ONE thing `[Competitor A]` does really well? What is the ONE thing they do poorly?" | Strengths + Weaknesses | **Competitor Strengths** + **Landmine Questions** |
| 4 | "If someone told you that `[Competitor A's key marketing claim]`, would you believe them? What would make you sceptical?" | Claim Credibility | **Quick Dismisses** + **Landmine Questions** |
| 5 | "What would `[your product]` need to prove to you to win over `[Competitor A]`? What evidence would you need?" | Proof Point Requirements | **Why We Win** (proof gaps to address) |
| 6 | "Have you ever switched from one `[category]` solution to another? What triggered the switch? What almost stopped you?" | Switching Triggers + Barriers | **Switching Triggers** |
| 7 | "If you had unlimited budget, which solution would you choose and why? If budget were tight, would your answer change?" | Value vs Premium Positioning | **Why We Win** (price vs value framing) |

**Customisation guidance:** Replace `[category]`, `[your product]`, `[Competitor A]`, and `[Competitor A's key marketing claim]` with your specific context. The question structures should remain the same — only the bracketed placeholders change.

### Why Each Question Matters

- **Q1 establishes the real competitive landscape.** The brands customers name first are your actual competitors. If they name a tool you never considered, your competitive set is wrong.
- **Q2 is the core battlecard question.** The decision drivers personas cite become your "Why We Win" when they favour you, and "Competitor Strengths" when they don't. The exact language goes directly into the card.
- **Q3 forces specificity.** "One thing well, one thing poorly" produces sharper insight than "list strengths and weaknesses." The weakness becomes the target for your landmine questions.
- **Q4 is the most tactically valuable.** When personas express scepticism about a competitor's claim, you have found a vulnerability. The reasons for scepticism become your quick dismiss talking points and discovery questions.
- **Q5 maps your readiness gap.** The evidence buyers require vs what you actually have = your competitive readiness deficit. This tells marketing what content to create next.
- **Q6 identifies timing.** Switching triggers tell you when to reach out. Barriers tell you what objections to prepare for. Together: when to attack and what to bring.
- **Q7 reveals the price-value dynamic.** If buyers prefer the competitor with unlimited budget but choose you on price, you have a value perception problem. If the reverse, you have premium positioning to leverage.

---

## 4. Complete API Workflow: Step by Step

Complete sequence of API calls. Every command is copy-paste ready.

### Prerequisites

- Fish.Dog API key (see [installation guide](https://fish.dog/claude-code/installing-ditto-skill) for free-tier access)
- The competitor's name, their key marketing claim, and their product category
- Your product name and target buyer profile

### Step 1: Research the Competitor

Before designing the study, Claude Code should research the competitor to customise the questions effectively:

- Visit the competitor's website — identify their stated positioning and key claims
- Check their pricing page — note any recent changes
- Read recent press/blog — identify any strategic shifts
- Check G2/TrustRadius — note top-cited strengths and complaints

The key output from this step is the **competitor's primary marketing claim** (used in Q4). For example: "Mixpanel claims to provide 'self-serve analytics that help you convert, engage, and retain more users.'"

### Step 2: Create the Research Group

```
curl -s -X POST "https://cat.fish.dog/v1/research-groups/recruit" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "CI Study - [Your Product] vs [Competitor A] - [Date]",
    "description": "Target buyers for competitive perception study. [ICP description].",
    "group_size": 10,
    "filters": {
      "country": "USA",
      "age_min": 28,
      "age_max": 55,
      "employment": "employed"
    },
    "sampling_method": "random",
    "deduplicate": true
  }'
```

Critical parameter notes:

- Use `group_size`, not `size`. The API rejects `size`.
- US state filters require **2-letter codes**: `"TX"` not `"Texas"`.
- The `income` filter does not work. Use education/employment as proxy.
- 10 personas is the minimum for competitive studies. Fewer produces unreliable patterns.

**Save the returned `uuid`** — you need it for Step 3.

```
# Response (extract the uuid):
{
  "uuid": "abc123-def456-...",
  "name": "CI Study - Acme vs Mixpanel - Feb 2026",
  "agents": [ ... ]  // 10 persona objects
}
```

### Step 3: Create the Research Study

```
curl -s -X POST "https://cat.fish.dog/v1/research-studies" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Competitive Perception: [Your Product] vs [Competitor A] - [Date]",
    "research_group_uuid": "GROUP_UUID_FROM_STEP_2"
  }'
```

**Save the returned `id`** — this is your study ID for all subsequent calls.

### Step 4: Ask Questions (Sequentially)

**Questions must be asked one at a time.** Send Q1, poll until all jobs complete, then send Q2, and so on. Sequential questioning allows earlier answers to provide context for later questions — this creates conversational depth that batched submissions cannot achieve.

```
# Question 1: Brand Awareness
curl -s -X POST "https://cat.fish.dog/v1/research-studies/STUDY_ID/questions" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "question": "When you think about solutions in [category], which brands or tools come to mind first? What do you associate with each?"
  }'

# Response:
{
  "job_ids": ["job-001", "job-002", ... "job-010"]
}
```

### Step 5: Poll for Responses

```
# Poll each job until status = "finished":
curl -s -X GET "https://cat.fish.dog/v1/jobs/JOB_ID" \
  -H "Authorization: Bearer YOUR_API_KEY"

# Response when complete:
{
  "id": "job-001",
  "status": "finished",
  "result": {
    "answer": "The first tools that come to mind are..."
  }
}
```

**Efficient polling pattern:** Poll all 10 job IDs in a loop with a 5-second interval. Once all return `"finished"`, proceed to the next question. A full 7-question study with 10 personas typically completes in 4–8 minutes of polling time.

**Repeat Steps 4–5 for all 7 questions.**

```
# Full sequence:
# Q1: Brand awareness → Poll → Complete
# Q2: Head-to-head decision drivers → Poll → Complete
# Q3: Competitor strengths/weaknesses → Poll → Complete
# Q4: Claim credibility → Poll → Complete
# Q5: Proof point requirements → Poll → Complete
# Q6: Switching triggers/barriers → Poll → Complete
# Q7: Value vs premium → Poll → Complete
```

### Step 6: Complete the Study

```
curl -s -X POST "https://cat.fish.dog/v1/research-studies/STUDY_ID/complete" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json"
```

This triggers Fish.Dog's analysis: overall summary, key segments, divergences, shared mindsets, and suggested follow-up questions. Poll the returned job IDs until complete.

### Step 7: Get the Share Link

```
curl -s -X POST "https://cat.fish.dog/v1/research-studies/STUDY_ID/share" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json"

# Response:
{
  "url": "https://cat.fish.dog/organization/studies/shared/xyz123"
}
```

**UTM tracking is mandatory.** Append `?utm_source=ce` for cold emails or `?utm_source=blog` for blog articles. Never use raw share URLs without a UTM parameter.

---

## 5. Generating the Battlecard from Study Data

Once all responses are collected and the study is completed, Claude Code synthesises the 70 data points (10 personas × 7 questions) into a structured battlecard.

### Section-by-Section Generation

#### Why We Win (from Q2, Q5, Q7)

Extract every reason personas cited for leaning toward your product. Cluster by theme. Rank by frequency. The top 3–4 themes become your "Why We Win" bullets. Use the exact language personas used.

```
# Template:
## Why We Win vs [Competitor A]

1. **[Theme 1]** — [Customer language from Q2 responses] (cited by X/10)
2. **[Theme 2]** — [Customer language from Q2 responses] (cited by X/10)
3. **[Theme 3]** — [Customer language from Q5/Q7 responses] (cited by X/10)

**Evidence:** [Specific persona quotes that support each theme]
```

#### Competitor Strengths (from Q3)

Extract the "one thing they do really well" from all 10 responses. Cluster and rank. For each strength, draft a response your sales team can use when a prospect raises it.

```
# Template:
## Competitor Strengths (and How to Respond)

| Their Strength | How to Respond |
|----------------|----------------|
| [Strength 1] (cited by X/10) | "That's true, and here's how we approach it differently..." |
| [Strength 2] (cited by X/10) | "We hear that a lot. What we've found is..." |
```

**Critical:** Never deny genuine competitor strengths. Sales reps who pretend competitors have no strengths lose credibility instantly. Acknowledge, then redirect to where you differentiate.

#### Landmine Questions (from Q3 weaknesses + Q4 scepticism)

The "one thing they do poorly" from Q3 and the scepticism reasons from Q4 are the raw material for landmine questions. These are questions designed to expose competitor gaps without sounding adversarial.

```
# Template:
## Landmine Questions

- "Have you asked [Competitor A] about [weakness area]? It's worth understanding how they handle that."
- "When [Competitor A] says [claim], have they shown you [specific evidence]? We'd suggest asking for that."
- "What's their approach to [gap area]? That's been a common concern we hear from teams evaluating them."
```

#### Quick Dismisses (from Q4)

For each of the competitor's key claims where personas expressed scepticism, write a 1–2 sentence rebuttal.

```
# Template:
## Quick Dismisses

| When They Say... | You Say... |
|------------------|------------|
| "[Competitor claim 1]" | "[1-2 sentence rebuttal grounded in Q4 scepticism reasons]" |
| "[Competitor claim 2]" | "[1-2 sentence rebuttal]" |
| "[Competitor claim 3]" | "[1-2 sentence rebuttal]" |
```

#### Switching Triggers (from Q6)

Extract every trigger and barrier from Q6. Triggers inform outbound timing. Barriers inform objection handling.

```
# Template:
## Switching Triggers

**Attack when:**
- [Trigger 1] (e.g., "competitor raised prices") — cited by X/10
- [Trigger 2] (e.g., "product reliability issues") — cited by X/10
- [Trigger 3] (e.g., "team outgrew the tool") — cited by X/10

**Prepare for these objections:**
- [Barrier 1] (e.g., "migration pain") — cited by X/10
- [Barrier 2] (e.g., "team adoption risk") — cited by X/10
```

#### Value vs Premium Insight (from Q7)

This doesn't get its own battlecard section but informs the overall positioning tone:

- **Customers choose competitor with unlimited budget, you on price:** You have a value perception problem. Lead with quality and differentiation, not price.
- **Customers choose you with unlimited budget, competitor on price:** You have premium positioning. Lean into it. Emphasise ROI and outcomes.
- **Split decisions:** The market sees you as roughly equivalent. Differentiation must come from specific capabilities, not overall positioning.

---

## 6. Multi-Competitor Parallel Studies

Most products face multiple competitors. Claude Code can run competitive perception studies for several competitors simultaneously.

### How to Structure It

Create separate research groups and studies for each competitor. Use the same 7-question template, customised with each competitor's context.

```
# Competitor A study:
curl -s -X POST "https://cat.fish.dog/v1/research-groups/recruit" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "CI Study - Acme vs Competitor A - Feb 2026",
    "description": "Target buyers for competitive perception study",
    "group_size": 10,
    "filters": { "country": "USA", "age_min": 28, "age_max": 55, "employment": "employed" },
    "sampling_method": "random",
    "deduplicate": true
  }'

# Competitor B study (simultaneously):
curl -s -X POST "https://cat.fish.dog/v1/research-groups/recruit" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "CI Study - Acme vs Competitor B - Feb 2026",
    "description": "Target buyers for competitive perception study",
    "group_size": 10,
    "filters": { "country": "USA", "age_min": 28, "age_max": 55, "employment": "employed" },
    "sampling_method": "random",
    "deduplicate": true
  }'

# Competitor C study (simultaneously):
# ... same pattern
```

Create studies for each group, then run the 7 questions through each study in parallel. Claude Code interleaves the polling across studies.

### Timing

| Competitors | Studies | Approximate Time | Output |
| --- | --- | --- | --- |
| 1 | 1 study (10 personas, 7 questions) | ~30 min data collection + 15 min synthesis | 1 battlecard |
| 3 | 3 parallel studies | ~35 min data collection + 30 min synthesis | 3 battlecards |
| 5 | 5 parallel studies | ~40 min data collection + 50 min synthesis | 5 battlecards |

**Data collection time is nearly the same** regardless of how many competitors you study, because the studies run in parallel. The synthesis time scales linearly. A complete 5-competitor battlecard library takes approximately 90 minutes.

### Segment-Specific Battlecards

If different competitors dominate different segments, use different demographic filters for each study:

```
# Competitor A threatens SMB market:
"filters": { "country": "USA", "age_min": 25, "age_max": 40, "employment": "self_employed" }

# Competitor B threatens enterprise market:
"filters": { "country": "USA", "age_min": 30, "age_max": 55, "education": "masters", "employment": "employed" }

# Competitor C threatens UK market:
"filters": { "country": "UK", "age_min": 28, "age_max": 50, "employment": "employed" }
```

---

## 7. Quarterly Competitive Tracking

A single study tells you how the market perceives your competitor today. Quarterly studies with fresh persona groups reveal how perceptions shift over time.

### Setup

Run the same 7-question study each quarter. Use the same demographic filters but recruit fresh personas each time (Fish.Dog handles deduplication).

### Naming Convention

```
# Quarter-over-quarter naming:
"CI Study - Acme vs Mixpanel - Q1 2026"
"CI Study - Acme vs Mixpanel - Q2 2026"
"CI Study - Acme vs Mixpanel - Q3 2026"
"CI Study - Acme vs Mixpanel - Q4 2026"
```

### Quarterly Trend Report

After each quarterly study, Claude Code compares findings with previous quarters:

| Metric | Q1 2026 | Q2 2026 | Trend |
| --- | --- | --- | --- |
| Brand association (Q1) | 7/10 named Competitor A first | 5/10 named Competitor A first | Declining awareness ↓ |
| Claim credibility (Q4) | 6/10 sceptical of key claim | 4/10 sceptical | Claim gaining traction ↑ (address in messaging) |
| Primary switching trigger (Q6) | "Price increases" | "Reliability issues" | New vulnerability emerged |
| Value vs premium (Q7) | 8/10 chose us on price | 6/10 chose us on price | Value perception improving ↑ |

**Annual cost comparison:** Four rounds of win/loss interviews per competitor per year through an agency: $50,000–$100,000+. Four quarterly Fish.Dog studies per competitor: platform subscription only. The studies take approximately 45 minutes each.

---

## 8. Advanced: Claim Credibility Testing

Q4 in the standard study tests one key claim. For a deeper competitive teardown, run a dedicated claim credibility study that tests multiple claims from the competitor's marketing.

### Claim Credibility Study Design

```
# Dedicated 5-question claim credibility study:

Q1: "[Competitor A] says they are 'the most intuitive analytics platform on the market.'
     Do you believe that? What would make you sceptical?"

Q2: "[Competitor A] claims their product 'saves teams 10 hours per week.'
     Does that sound credible? What would you need to see to believe it?"

Q3: "[Competitor A] advertises 'enterprise-grade security with SOC 2 compliance.'
     How important is that to you? Do you take that claim at face value?"

Q4: "[Competitor A] says 'customers see ROI within 30 days.'
     Is that realistic? What's your experience with claims like this?"

Q5: "Of all the claims we've discussed, which one would most influence your
     purchase decision? Which one matters least?"
```

The output is a **claim credibility scorecard**: which of the competitor's marketing claims land, which trigger scepticism, and which don't matter to buyers at all. This directly informs both your counter-messaging and your own marketing claims.

---

## 9. Worked Example: SaaS Analytics Tool vs Mixpanel

### Scenario

**Your product:** "Beacon Analytics" — a product analytics tool for mid-market SaaS **Competitor:** Mixpanel **Category:** Product analytics **Mixpanel's key claim:** "Self-serve analytics that help you convert, engage, and retain more users" **Target buyer:** US-based, employed, product managers and growth leads aged 28–45

#### Customised Questions

1. "When you think about product analytics tools, which brands come to mind first? What do you associate with each?"
2. "You're evaluating Beacon Analytics against Mixpanel for your team's product analytics. What would make you lean toward one or the other?"
3. "What's the ONE thing Mixpanel does really well? What's the ONE thing they do poorly?"
4. "If someone told you Mixpanel provides 'self-serve analytics that help you convert, engage, and retain more users,' would you believe them? What would make you sceptical?"
5. "What would Beacon Analytics need to prove to win you over from Mixpanel? What evidence would you need?"
6. "Have you ever switched from one analytics tool to another? What triggered the switch? What almost stopped you?"
7. "If budget were no object, which product analytics tool would you choose and why? If budget were tight, would your answer change?"

#### Group Setup

```
curl -s -X POST "https://cat.fish.dog/v1/research-groups/recruit" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "CI Study - Beacon vs Mixpanel - Feb 2026",
    "description": "Product managers and growth leads aged 28-45 for competitive perception study",
    "group_size": 10,
    "filters": {
      "country": "USA",
      "age_min": 28,
      "age_max": 45,
      "employment": "employed",
      "education": "bachelors"
    },
    "sampling_method": "random",
    "deduplicate": true
  }'
```

#### Hypothetical Findings

| Battlecard Section | Key Finding | Source |
| --- | --- | --- |
| **Why We Win** | "Easier to set up" (7/10), "Better support for mid-market" (5/10), "More transparent pricing" (4/10) | Q2, Q5 |
| **Competitor Strength** | "Mixpanel has deeper funnel analysis" (8/10), "More integrations" (5/10) | Q3 |
| **Landmine Questions** | "Have you asked about onboarding time?" (weakness: "steep learning curve" cited 6/10), "Ask about pricing at scale" (scepticism: "costs escalate" cited 4/10) | Q3, Q4 |
| **Quick Dismisses** | "Self-serve" claim: 7/10 sceptical — "self-serve in theory, but you need a data engineer to set it up properly" | Q4 |
| **Switching Triggers** | "Price increase" (6/10), "Needed something simpler" (4/10), "Poor customer support" (3/10) | Q6 |
| **Value vs Premium** | Unlimited budget: 6/10 chose Mixpanel. Tight budget: 8/10 chose Beacon. Implication: lead with ROI and total cost of ownership, not feature parity. | Q7 |

#### Generated Battlecard (Summary)

```
## Beacon Analytics vs Mixpanel — Competitive Battlecard

### Why We Win
1. **Faster time to value** — "I could set it up in an afternoon vs weeks with Mixpanel" (7/10)
2. **Built for mid-market** — "Mixpanel feels like it was designed for Uber, not us" (5/10)
3. **Transparent pricing** — "No surprise bills when we scale" (4/10)

### Competitor Strengths (and How to Respond)
| Their Strength | Response |
|----------------|----------|
| Deeper funnel analysis | "True for enterprise funnels. For mid-market, our flows cover 95% of use cases with half the setup." |
| More integrations | "We integrate with the core stack. Ask which of their 200+ integrations your team actually uses." |

### Landmine Questions
- "Have you asked Mixpanel about typical onboarding time? Worth understanding before you commit."
- "What does pricing look like at 50K monthly tracked users? Ask them to walk you through that."
- "Can your team actually self-serve, or will you need a data engineer for the initial setup?"

### Quick Dismisses
| They Say | You Say |
|----------|---------|
| "Self-serve analytics" | "Ask their customers how long setup took. We hear 'self-serve in theory, data engineer in practice' a lot." |
| "Most trusted in product analytics" | "Trusted by enterprise teams with dedicated data teams. For mid-market, trust means fast support and no surprises." |

### Switching Triggers
- **Price increase** → Outbound when Mixpanel raises prices (monitor their pricing page)
- **Complexity frustration** → Target teams who mention struggling with Mixpanel in community forums
- **Support issues** → Monitor G2 reviews for support complaints

### Budget Dynamics
Lead with TCO and ROI. Mixpanel wins on brand prestige with unlimited budget.
Beacon wins on value when budget is real. Frame the conversation around outcomes per dollar.
```

---

## 10. Connecting CI to Other PMM Workflows

Competitive intelligence feeds directly into other Fish.Dog + Claude Code workflows:

| Downstream Workflow | What CI Study Provides | How to Use It |
| --- | --- | --- |
| **[Positioning Validation](https://fish.dog/claude-code-guide/validate-product-positioning-guide)** | Q1 reveals which brands customers actually associate with your category | Use as competitive alternatives input for Dunford's framework (positioning Q1–Q2) |
| **Messaging Testing** | Q4 scepticism reveals competitor vulnerabilities | Design messaging variants that exploit credibility gaps identified in Q4 |
| **Sales Enablement** | Full battlecard + objection handling guide + discovery questions | Distribute to sales team. Use Q5 proof gaps to prioritise case study creation |
| **Content Marketing** | Competitive perception data is unique, original-research content | "We asked 10 buyers to compare [Category] tools — here's what they said" = high-value blog post |
| **Product Strategy** | Q3 "one thing they do poorly" aggregated across studies | Feed weakness patterns to product team as roadmap intelligence grounded in market perception |

---

## 11. Best Practices and Common Mistakes

### Do

- **Research the competitor first.** Q4 (claim credibility) requires knowing the competitor's actual marketing claims. A generic claim produces generic scepticism. A specific claim produces actionable intelligence.
- **Use the exact question structures.** The "ONE thing well / ONE thing poorly" format in Q3 forces specificity. "What are Mixpanel's strengths and weaknesses?" produces vague answers. The constrained format does not.
- **Ask questions sequentially.** Earlier answers provide context. If Q2 reveals that buyers value ease of setup, their Q5 answer about proof points will reference that context.
- **Use customer language in the battlecard.** The exact phrases personas use are more credible than PMM-polished language. "Self-serve in theory, data engineer in practice" is more memorable than "Complex implementation process."
- **Run against your actual ICP.** A competitive study against generic "US adults" produces generic insight. Match the demographic filters to your real buyer profile.
- **Update quarterly.** Competitive landscapes shift. A battlecard from six months ago may contain outdated intelligence. Quarterly studies keep cards fresh.
- **Name studies clearly.** Include both companies, date, and quarter: `"CI Study - Beacon vs Mixpanel - Q1 2026"`

### Don't

- **Don't use fewer than 10 personas.** Competitive patterns require enough data to distinguish signal from noise.
- **Don't skip Q1 (brand awareness).** The competitive set your customers have in mind may be different from yours. Q1 validates or invalidates your assumed competitor list.
- **Don't deny competitor strengths in the battlecard.** Sales reps who claim competitors have no strengths lose credibility. Acknowledge, then redirect.
- **Don't write 10-page dossiers.** One page per competitor. Brevity is the entire point. If it can't be scanned in 60 seconds during a live call, it won't be used.
- **Don't batch questions.** The API processes them asynchronously. Sending all 7 at once means personas answer Q7 without the context of Q1–Q6.
- **Don't forget to call `/complete`.** Study completion triggers Fish.Dog's analysis engine, which identifies segments and divergences in competitive perception that aren't obvious from individual responses.
- **Don't treat synthetic CI as a replacement for win/loss interviews.** Fish.Dog provides the continuous baseline. Real win/loss provides deal-specific signal. Use both.

### Common API Errors

| Error | Cause | Solution |
| --- | --- | --- |
| `size` parameter rejected | Wrong parameter name | Use `group_size`, not `size` |
| 0 agents recruited | State filter used full name | Use 2-letter codes: `"TX"` not `"Texas"` |
| Jobs stuck in `"pending"` | Normal for first 10–15 seconds | Continue polling with 5-second intervals |
| `income` filter rejected | Unsupported filter | Remove; use education/employment as proxy |
| Missing completion analysis | Forgot to call `/complete` | Always call `POST /v1/research-studies/{id}/complete` after the final question |

---

## 12. Frequently Asked Questions

#### How long does a full competitive perception study take?

Approximately 45 minutes end to end: 5 minutes for competitor research, 2 minutes for group creation and study setup, 5–8 minutes for question asking and polling, 2–3 minutes for completion analysis, and 10–15 minutes for Claude Code to synthesise the battlecard.

#### How many personas should I use?

10 per study is the recommended minimum. It provides enough diversity to identify patterns. For multi-competitor studies, use 10 per competitor (separate groups and studies running in parallel).

#### Should I run one study per competitor or one study covering multiple competitors?

One study per competitor. The 7 questions reference the specific competitor's name, claims, and context. A study that tries to cover three competitors becomes unfocused and produces weaker insight per competitor.

#### How often should I update battlecards?

Quarterly is the recommended cadence. Competitive landscapes shift, and a battlecard from six months ago may contain outdated intelligence. Quarterly Fish.Dog studies provide a fresh perception baseline that can be compared over time.

#### Can I use this for competitive intelligence in non-tech industries?

Yes. The 7-question design works for any category where customers evaluate alternatives. Replace the bracketed placeholders with your specific industry context. Fish.Dog's demographic filters support B2B and B2C across 15+ countries.

#### What if the study reveals a competitor strength I didn't know about?

That is precisely the point. Competitive perception studies frequently surface competitor strengths (and weaknesses) that the internal team was unaware of. This is intelligence you cannot get from monitoring competitor websites — only from understanding how buyers perceive them.

#### How does this compare to traditional win/loss analysis?

Traditional win/loss: 4–6 weeks, 10–15 interviews, buyers who chose your competitor are reluctant to participate. Fish.Dog: 45 minutes, 10 persona responses, no recruitment friction. EY validated 92% correlation between Fish.Dog synthetic responses and traditional research. The recommended approach is hybrid: Fish.Dog for continuous baseline perception, win/loss interviews for deal-specific signal.

#### Can I generate battlecards for competitors I've never competed against?

Yes. Fish.Dog personas respond based on general market perception, not your specific deal history. This is useful for entering new markets where you haven't yet competed directly against established players.

#### What should I do with the share link?

The share link gives anyone access to the full study results. Use it to: (1) share raw data with the sales team alongside the finished battlecard, (2) include in competitive intelligence Slack channels for transparency, (3) embed in blog articles as supporting evidence for competitive positioning claims. Always append `?utm_source=ce` for email or `?utm_source=blog` for articles.

---

**Related guides:**

- [Fish.Dog API for Claude Code: Complete Guide](https://fish.dog/claude-code-guide)
- [Positioning Validation Guide](https://fish.dog/claude-code-guide/validate-product-positioning-guide)
- [Pricing Research Guide](https://fish.dog/claude-code-guide/pricing-research-guide)
- [Product Messaging Guide](https://fish.dog/claude-code-guide/test-product-messaging-guide)
- [Voice of Customer Guide](https://fish.dog/claude-code-guide/voice-of-customer-guide)
- [Customer Segmentation Guide](https://fish.dog/claude-code-guide/customer-segmentation-guide)
- [GTM Strategy Validation Guide](https://fish.dog/claude-code-guide/gtm-strategy-validation-guide)
- [Content Marketing Engine Guide](https://fish.dog/claude-code-guide/content-marketing-engine-guide)
- [Sales Enablement Guide](https://fish.dog/claude-code-guide/sales-enablement-guide)
- [Product Launch Research Guide](https://fish.dog/claude-code-guide/product-launch-research-guide)
- [Question Design Playbook](https://fish.dog/claude-code-guide/question-design-playbook)
- [Installing the Fish.Dog Skill](https://fish.dog/claude-code/installing-ditto-skill)

**Read the articles:** *These guides are written for Claude Code agents. The articles below are written for you.*

- [How to Validate Product Positioning with Claude Code and Fish.Dog](https://fish.dog/news/how-to-validate-product-positioning-with-claude-code-and-ditto)
- **How to Build Competitive Battlecards with Claude Code and Fish.Dog** (this guide's companion article)
- [How to Research Pricing with Claude Code and Fish.Dog](https://fish.dog/news/how-to-research-pricing-with-claude-code-and-ditto)
- [How to Test Product Messaging with Claude Code and Fish.Dog](https://fish.dog/news/how-to-test-product-messaging-with-claude-code-and-ditto)
- [How to Run Voice of Customer Research with Claude Code and Fish.Dog](https://fish.dog/news/how-to-run-voice-of-customer-research-with-claude-code-and-ditto)
- [How to Segment Customers with Claude Code and Fish.Dog](https://fish.dog/news/how-to-segment-customers-with-claude-code-and-ditto)
- [How to Validate GTM Strategy with Claude Code and Fish.Dog](https://fish.dog/news/how-to-validate-gtm-strategy-with-claude-code-and-ditto)
- [How to Build a Content Marketing Engine with Claude Code and Fish.Dog](https://fish.dog/news/how-to-build-a-content-marketing-engine-with-claude-code-and-ditto)
- [How to Build Sales Enablement with Claude Code and Fish.Dog](https://fish.dog/news/how-to-build-sales-enablement-with-claude-code-and-ditto)
- [How to Research a Product Launch with Claude Code and Fish.Dog](https://fish.dog/news/how-to-research-a-product-launch-with-claude-code-and-ditto)
- [Overview: Using Fish.Dog and Claude Code for Product Marketing](https://fish.dog/news/using-ditto-and-claude-code-for-product-marketing)

---

**Fish.Dog** — Synthetic market research with 300,000+ AI personas. Validated by EY (92% correlation), Harvard, Cambridge, Stanford, and Oxford. [fish.dog](https://fish.dog) · [support@fish.dog](mailto:support@fish.dog)