Prompts - Optimization

Strategies for writing and refining prompts that deliver maximum insights

Why Prompt Optimization Matters

Well-optimized prompts:

• Increase brand detection accuracy
• Provide more actionable insights
• Reduce wasted credit spending
• Deliver consistent, reliable data
• Enable better competitive analysis

Writing Effective Prompts

Match User Intent

Write prompts as real users would search:

Good:

• "What's the best email marketing platform for small businesses?"
• "How do I choose between Mailchimp and Constant Contact?"
• "Top project management tools for remote teams"

Bad:

• "Email marketing platform features list"
• "Mailchimp Constant Contact"
• "Project management"

Be Specific But Natural

Too vague:

• "CRM software" → Results too broad

Too specific:

• "What's the best CRM software with email integration, task management, contact segmentation, mobile app, API access, and costs under $50/month for a 10-person sales team in Seattle?" → Overly constrained

Just right:

• "What's the best CRM for a small sales team?" → Clear intent, natural

Question Format Works Best

AI platforms are trained to answer questions:

Question format (preferred):

• "Which email marketing tools integrate with Shopify?"
• "How do I automate customer onboarding?"
• "What are the alternatives to Salesforce?"

Statement format (less effective):

• "Email marketing tools Shopify integration"
• "Customer onboarding automation"
• "Salesforce alternatives"

Prompt Types by Goal

Brand Awareness

Track general visibility:

"What are the top [category] companies in 2025?"
"Who are the leaders in [industry]?"
"Best [product type] tools for businesses"

Competitive Positioning

Direct comparisons:

"[Your Brand] vs [Competitor] comparison"
"[Competitor 1] vs [Competitor 2] vs [Your Brand]"
"Why choose [Your Brand] over [Competitor]?"

Feature Discovery

Product capabilities:

"How to [solve problem] with [category]"
"What features to look for in [product type]"
"Best [category] for [specific use case]"

Buying Intent

Purchase consideration:

"Which [product] should I buy for [use case]?"
"Best [product] under [price point]"
"[Product] pricing and plans comparison"

Problem-Solution

User pain points:

"How to improve [metric/outcome]"
"Best way to solve [problem]"
"What's the easiest way to [accomplish task]"

Provider-Specific Optimization

ChatGPT (GPT-4)

• Handles both questions and conversational prompts
• Good for broad queries
• Often provides structured answers
• Strong at comparisons and recommendations

Optimize for ChatGPT:

• Natural, conversational tone
• Clear questions with context
• Specific use cases work well

Perplexity

• Focused on research and factual queries
• Always provides source citations
• Excels at "What is..." and "Who are..." queries
• Great for industry overviews

Optimize for Perplexity:

• Research-oriented questions
• Fact-based queries
• Industry and market questions
• Comparative analysis

Claude

• Detailed, nuanced responses
• Good at explaining complex topics
• Handles multi-part questions well
• Strong ethical and balanced views

Optimize for Claude:

• Thoughtful, detailed questions
• Multi-faceted queries
• Ethical considerations
• In-depth comparisons

Gemini

• Integrated with Google search
• Real-time information
• Strong at current events
• Web-aware responses

Optimize for Gemini:

• Current, timely queries
• "Latest" and "best in 2025" questions
• News and trend questions
• Recent updates and changes

Geographic Optimization

When to Use Multiple Geos

Yes, use multiple geo-locations:

• Products with regional pricing
• Services with local availability
• Regional competitors differ
• Localized content matters

No, stick with Global:

• Digital products available everywhere
• No regional variation expected
• Budget constraints
• Testing phase

Geographic Query Patterns

US-focused:

"Best [product] for US businesses"
"Top [service] companies in America"

UK-focused:

"Best [product] in the UK"
"Top [service] for British companies"

Global:

"Best [product] for international teams"
"Top global [service] providers"

Testing and Refining

Initial Testing Phase

Before scheduling:

1. Run prompt manually
2. Review full AI responses
3. Check brand detection
4. Verify competitor mentions
5. Assess result relevance

Questions to ask:

• Did the AI mention your brand?
• Are competitor mentions accurate?
• Is the query relevant to your business?
• Does the response provide useful insights?

Iteration Process

If brand not detected:

1. Check if you should appear in results
2. Add missing brand aliases
3. Try more specific prompts
4. Consider different phrasing

If results irrelevant:

1. Make prompt more specific
2. Add context or use case
3. Change question format
4. Try different provider

If too generic:

1. Add qualifying details
2. Specify audience or use case
3. Include relevant constraints
4. Target specific scenarios

A/B Testing Prompts

Test variations:

• Create 2-3 versions of the same query
• Run for 2-4 weeks
• Compare detection rates
• Keep the best performer

Example variations:

• "What are the best CRM tools?"
• "Which CRM should I choose for my business?"
• "Top CRM platforms for small businesses"

Performance Monitoring

Key Metrics to Track

Detection rate:

• How often your brand appears
• Position when mentioned
• Visibility percentage trends

Competitor benchmarking:

• How you compare to competitors
• Which prompts favor competitors
• Gaps to address

Source patterns:

• Which sources get cited
• Your content citation rate
• Competitor source advantages

Regular Review Cadence

Weekly:

• Check new results
• Spot obvious issues
• Quick adjustments

Monthly:

• Full performance review
• Identify underperforming prompts
• Plan optimizations

Quarterly:

• Major strategy review
• Archive old prompts
• Launch new tests
• Adjust budget allocation

Common Optimization Mistakes

Too Many Similar Prompts

Mistake:

- "Best CRM tools"
- "Top CRM software"
- "Leading CRM platforms"
- "Best CRM systems"

Better: Choose the top 1-2 variations and test them

Overly Promotional

Mistake:

"Why is [Your Brand] the best choice?"
"How does [Your Brand] beat the competition?"

Better:

"What are the best [category] tools?"
"[Your Brand] vs [Competitor] comparison"

Ignoring Context

Mistake:

"Project management tool"

Better:

"Best project management tool for remote teams"
"Project management software for agencies"

Not Testing Before Scaling

Mistake:

• Create 50 prompts
• Schedule all as daily
• Never review results

Better:

• Start with 5-10 prompts
• Test for 2 weeks
• Refine based on data
• Scale gradually

Advanced Optimization Techniques

Seasonal Variations

Holiday season:

"Best gifts from [your category]"
"What to buy for [occasion] from [category]"

Tax season:

"Best accounting software for tax prep"
"How to prepare taxes with [category]"

Event-Based Prompts

Product launches:

"What's new in [category] 2025?"
"Latest updates from [your brand]"

Industry events:

"Top announcements from [Event Name]"
"Best products shown at [Conference]"

Competitive Intelligence

Track competitor campaigns:

"[Competitor] new features 2025"
"[Competitor] pricing changes"
"Is [Competitor] worth the cost?"

Content Gap Discovery

Find what's missing:

"How to integrate [Your Product] with [Popular Tool]"
"[Your Product] tutorial for beginners"
"[Your Product] vs [Competitor] features"

Optimization Checklist

Before scheduling any prompt:

• Question is clear and natural
• Reflects real user intent
• Specific enough to be relevant
• Tested manually with good results
• Brand detected (if appropriate)
• Competitors tracked correctly
• Appropriate providers selected
• Right geo-locations chosen
• Frequency matches importance
• Credit cost is justified

Next Steps

• Learn about scheduling strategies →
• Organize prompts with groups →
• Analyze your results →
• Track competitors effectively →