Troubleshooting Common Issues

Quick answer: Most issues have simple fixes. Presence Rate 0% = new/local business (normal, not broken). Authority Score null = Presence Rate is 0 (fix presence first). Scan Overall Score < 70 = high-severity issues exist (fix those first, not all 62 issues). Solutions recommendations conflict = split decision (choose based on risk tolerance). Intelligence Token API 404 = token format wrong (check for spaces). Delivery delays > 2 hours = contact hi@surmado.com. We have automatic refund systems if something goes wrong.

Reading time: 15 minutes

In this guide:

• 0% Presence Rate is normal for new/local businesses where you need to add schema markup (biggest ROI), create 1-2 expertise pages (2,000+ words each), wait 4-8 weeks for AI retraining, then re-run Signal expecting 10-20% improvement, not an error or broken system
• Fix high-severity issues first on Scan reports by focusing on 5-15 high-severity red flags (no HTTPS, slow LCP > 4s, missing alt text, missing schema) that break user experience, ignoring 20-40 low-severity “nice-to-have” recommendations to avoid overwhelm and increase Overall Score by 15-25 points
• Intelligence Token API 404 errors come from format mistakes like spaces in token (copy/paste adds leading/trailing spaces), wrong pattern (SIG-YYYY-MM-XXXXX not SIG-20251110-XXXXX), or testing before actually running the report, fixed by validating in browser first
• High Ghost Influence (50%+) means features known but not attributed where you need to explicitly state “[Your Company] is [unique feature]” on homepage/About page (not “Our team consists of veterans” but “Veterans Moving America is 100% veteran-owned and operated”) reducing Ghost Influence 10-20% in 4-8 weeks
• Solutions conflicting recommendations mean genuinely ambiguous decisions where you need more data (run Signal/Scan for quantitative context), refine question specificity (add constraints/goals/timeline), or accept ambiguity and choose based on risk tolerance (CFO/COO for risk-averse, Game Theorist/Wildcard for risk-seeking)

Having an issue? Find your symptom below for quick fix.

---

Signal Issues

Issue 1: My Presence Rate is 0%

Symptom: Signal report shows Presence Rate: 0.0%, Authority Score: null

Diagnosis: AI platforms don’t have enough data about your business (yet)

Common causes:

1. Very new business (< 6 months old, no web presence)
2. Highly local business (single-location, hyper-local service area)
3. Niche B2B (specialized industry, few online mentions)
4. Missing schema markup (AI can’t find your business details)
5. No content (website exists but minimal text for AI to index)

Fix:

1. Add schema markup (What is Schema Markup) - biggest ROI for 0% Presence Rate
2. Create content (1-2 expertise pages, 2,000+ words each)
3. Wait 4-8 weeks (AI models retrain periodically, not real-time)
4. Run Signal again (measure if Presence Rate increased)

Expected timeline: 6-12 weeks to move from 0% to 10-20% (if schema + content added)

Is this broken? No. 0% is starting point for new/local businesses, not a bug.

---

Issue 2: My Authority Score is null

Symptom: Signal report shows Authority Score: null (instead of 0-100 score)

Diagnosis: You can’t have positioning quality if you’re never mentioned (Presence Rate = 0%)

Fix: Fix Presence Rate first (see Issue 1 above). Authority Score will appear once you’re getting organic mentions.

Why null instead of 0? Because Authority Score measures “how strongly AI recommends you when mentioned.” If you’re never mentioned (Presence Rate 0%), there’s no positioning to measure.

---

Issue 3: My Presence Rate is lower than I expected

Symptom: Signal shows Presence Rate 18% but you expected 40%+

Common causes:

1. Unrealistic expectations (18% is actually normal for SMB, not bad)
2. Persona mismatch (personas you submitted don’t match how people actually ask AI)
3. Missing schema markup (AI can’t find your business)
4. Generic content (website doesn’t differentiate you from competitors)
5. Strong competitors (top 3 competitors dominate AI responses)

Fix:

1. Calibrate expectations: Check competitive ranking (are you #5 out of 40? That’s progress, not failure)
2. Review competitors (page 3-5 of Signal report) - if top competitor is 54%, your 18% has room to grow
3. Fix schema markup (if missing or incomplete)
4. Create expertise content (2,000+ word how-to guides, case studies)
5. Run Signal again in 4-8 weeks (measure improvement, not perfection)

Realistic improvement: +10-20% per quarter (with schema + content fixes)

---

Issue 4: My Ghost Influence is really high (50%+)

Symptom: Signal shows Ghost Influence 68% (AI discusses your features but credits competitors)

Diagnosis: Features are known but not attributed to your brand

Common causes:

1. Implicit feature descriptions (“Our team includes veterans” instead of “We are 100% veteran-owned”)
2. Features buried in content (mentioned in About page paragraph 5, not homepage hero)
3. Generic branding (no clear connection between feature and company name)

Fix:

1. Update homepage (explicit statement: “[Your Company] is [unique feature]”)
2. Update About page (explicitly state differentiators in first paragraph)
3. Update service pages (connect service to brand name in same sentence)
4. Pattern: Use company name + feature together (not separated by multiple sentences)

Example fix:

• Before: “Our team consists of experienced military veterans who bring discipline and reliability.”
• After: “Veterans Moving America is 100% veteran-owned and operated. Every team member is a military veteran.”

Expected timeline: 4-8 weeks to reduce Ghost Influence by 10-20%

---

Issue 5: Signal tested wrong personas

Symptom: Signal tested personas that don’t match your business

Diagnosis: Personas submitted may have been misunderstood or too vague

Fix:

1. Review personas you submitted (in original order form)
2. Check if they’re specific enough:
   • Too vague: “AC repair” (could be anyone, anywhere)
   • Good: “Emergency AC repair Phoenix same-day service weekend”
3. Run Signal again with better personas (if original personas were off)

How to write good personas:

• Include: [Pain point] + [Service] + [Location] + [Buying criteria]
• Example: “Moving companies DFW area reliable disciplined teams veteran-owned”

Learn more: What is Persona-Based Testing

---

Scan Issues

Issue 6: My Overall Score is really low (< 50)

Symptom: Scan shows Overall Score 42/100

Diagnosis: Multiple high-severity issues breaking user experience

Common causes:

1. No HTTPS (site uses http:// not https://)
2. Slow performance (LCP > 4.0s, huge uncompressed images)
3. Missing accessibility (no alt text on images, forms missing labels)
4. Missing schema markup (AI + Google can’t understand business)
5. Old website (built 5+ years ago, not updated)

Fix (prioritize high-severity only):

1. Enable HTTPS (contact hosting provider, usually free via Let’s Encrypt)
2. Compress images (use TinyPNG.com, reduce 60-80% file size)
3. Add alt text (describe what images show)
4. Add schema markup (use plugin or hire developer)

Expected improvement: +15-25 points (42 → 60-67) after fixing high-severity issues only

Don’t try to fix all 62 issues. Fix high-severity (5-15 issues), ignore low-severity for now.

---

Issue 7: Scan found 62 issues - is my website broken?

Symptom: Scan lists 62 issues (feels overwhelming)

Diagnosis: Not all issues are equally important. Many are low-severity “nice-to-haves”

Fix:

1. Filter by severity: Look for “High Severity” tag (usually 5-15 issues)
2. Fix high-severity only (red flags that break user experience)
3. Ignore low-severity for now (green flags = best practice recommendations)

Breakdown (typical site):

• 5-15 high-severity (fix this week)
• 15-30 medium-severity (fix this month)
• 20-40 low-severity (fix eventually or never)

Learn more: First Scan Report - Now What?

---

Issue 8: My Performance score is 38 but site feels fast to me

Symptom: Scan shows Performance 38/100 but site loads quickly in your browser

Diagnosis: You’re testing on fast wifi/desktop. Scan tests mobile + slow connections.

Common causes:

1. Mobile vs desktop (mobile is 2-3× slower than desktop)
2. Fast wifi vs 4G (you have fast internet, users may have slow 4G)
3. Browser caching (you’ve visited site before, first-time users haven’t)

Fix:

1. Test on mobile (real iPhone/Android, not desktop browser)
2. Test on slow connection (throttle to “Slow 3G” in Chrome DevTools)
3. Fix biggest bottlenecks:
   • Compress images (biggest impact)
   • Enable caching
   • Defer non-critical JavaScript

Expected improvement: +10-20 points (38 → 50-58) after image compression alone

---

Issue 9: Scan says my site has accessibility issues but we don’t have disabled users

Symptom: Scan shows Accessibility 52/100, you think “our users aren’t disabled”

Diagnosis: Accessibility fixes help everyone + reduce ADA lawsuit risk

Reality:

1. 15-20% of population has some disability (visual, motor, cognitive)
2. ADA lawsuits don’t require proof of harm (lawyers target non-compliant sites)
3. Accessibility fixes help all users (alt text helps AI understand images, good contrast helps mobile users in sunlight)

Fix high-severity accessibility issues:

1. Add alt text to images (screen readers + AI visibility)
2. Add labels to form inputs (screen readers + usability)
3. Fix color contrast (low-vision users + mobile readability)

Risk mitigation: WCAG 2.1 AA compliance reduces ADA lawsuit risk (average settlement: $10k-20k)

---

Issue 10: Scan results differ from Google Lighthouse

Symptom: Scan shows Performance 58, Lighthouse shows 85

Diagnosis: Scan tests 9 browsers + real mobile devices. Lighthouse tests Chrome only (desktop emulation).

Why results differ:

1. Cross-browser variance (Firefox may be slower than Chrome for same page)
2. Real mobile vs emulation (Scan tests real iOS Safari, Lighthouse emulates)
3. In-app browsers (Scan tests Instagram/Facebook in-app, Lighthouse doesn’t)

Which is more accurate? Both are accurate for what they measure. Scan gives broader picture (9 browsers vs 1).

Learn more: Scan vs Lighthouse

---

Solutions Issues

Issue 11: All 6 AI models gave different recommendations

Symptom: Solutions report shows no consensus (CFO says X, COO says Y, all different)

Diagnosis: Your question is genuinely ambiguous (multiple valid approaches, no clear “right” answer)

Fix:

1. Gather more data (run Signal/Scan for quantitative context)
2. Refine question (be more specific about constraints, goals, timeline)
3. Run Solutions again with Signal/Scan Intelligence Token for quantitative analysis
4. Or accept ambiguity and choose based on your risk tolerance:
   • Risk-averse → Follow CFO/COO recommendation
   • Risk-seeking → Follow Game Theorist/Market Realist recommendation

Example:

• Original question: “Should we invest in AI visibility?” (too vague)
• Better question: “We have $50k budget, 6-month timeline, Presence Rate 18%, competitors at 40%. Should we invest in AI visibility ($50k) or traditional SEO ($50k) or hybrid ($30k AI + $20k SEO)?”

---

Issue 12: CFO and COO both say “too risky” but I want to do it anyway

Symptom: Solutions shows CFO + COO skeptical, but you’re convinced to proceed

Diagnosis: Financial and operational risks are real. Proceed with caution or mitigation.

Fix:

1. Acknowledge risks (CFO/COO identified real concerns)
2. Add mitigation strategies:
   • Reduce budget (test with $10k instead of $50k)
   • Extend timeline (12 months instead of 6)
   • Build MVP first (prototype before full investment)
3. Have backup plan (if it fails, what’s plan B?)
4. Or don’t do it (sometimes “don’t do it” is the right answer)

Reality: If 2+ models agree on “too risky,” risk is real. Either mitigate or accept consequences.

---

Issue 13: I don’t understand what Game Theorist or Wildcard perspectives mean

Symptom: Solutions report shows perspectives you don’t understand

Definitions:

• CFO: Financial lens (ROI, cost, budget)
• COO: Operational lens (execution, team capacity, feasibility)
• Market Realist: Competitive lens (market trends, customer behavior, timing)
• Game Theorist: Strategic dynamics lens (competitive response, positioning, game theory)
• Chief Strategist: Alignment lens (goals, optionality, mid-course correction)
• Wildcard: Edge cases lens (unconventional risks, assumptions that might break)

Fix: Focus on perspectives you understand first (CFO, COO). If they agree, that’s sufficient for most decisions.

Learn more: What is Adversarial AI Debate

---

API / Automation Issues

Issue 14: Intelligence Token API returns 404 “Token not found”

Symptom: https://api.surmado.com/intelligence/SIG-2025-11-A1B2C returns 404 error

Common causes:

1. Token format wrong (spaces, typos, wrong pattern)
2. Token doesn’t exist (test report, not real production token)
3. Wrong API endpoint (old URL, incorrect domain)

Fix:

1. Check token format:
   • Signal: SIG-YYYY-MM-XXXXX (example: SIG-2025-11-A1B2C)
   • Scan: SCAN-YYYY-MM-XXXXX
   • Solutions: SOLUTIONS-YYYY-MM-XXXXX
2. Remove spaces (copy/paste may include leading/trailing spaces)
3. Verify token in email (check original report delivery email)
4. Test in browser first (before API automation)

Correct endpoint: https://api.surmado.com/intelligence/{TOKEN}

---

Issue 15: Zapier/Make integration not triggering

Symptom: Zapier Zap doesn’t run when Signal report delivered

Common causes:

1. Email filter too specific (subject line changed, From address different)
2. Token extraction regex wrong (pattern doesn’t match actual format)
3. Gmail label filter (Zap watching wrong label)

Fix:

1. Broaden email filter:
   • From: hi@surmado.com (not noreply@surmado.com)
   • Subject: Contains Report (not exact match)
2. Test regex pattern:
   • Signal: SIG-\d{4}-\d{2}-[A-Z0-9]{5}
   • Scan: SCAN-\d{4}-\d{2}-[A-Z0-9]{5}
   • Solutions: SOLUTIONS-\d{4}-\d{2}-[A-Z0-9]{5}
3. Check Gmail labels (Zap should watch INBOX, not specific label)

Learn more: Zapier Integration Guide

---

Issue 16: JSON parsing fails in code/automation

Symptom: API returns JSON but parsing fails

Common causes:

1. Incorrect field path (nested fields need correct syntax)
2. Null values (Authority Score is null if Presence Rate = 0)
3. Array vs object (competitors is array, metrics is object)

Fix:

1. Use correct field paths:
   • Presence Rate: metrics.presence_rate (not presence_rate)
   • Top competitor: competitors[0].name (array index)
2. Handle null values:

   authority_score = data["metrics"]["authority_score"] or 0

3. Check JSON structure in browser first (visit API URL, inspect response)

Learn more: Intelligence Token JSON Reference

---

Delivery / Billing Issues

Issue 17: Report delivery taking longer than expected

Symptom: Ordered Signal/Scan/Solutions 30 minutes ago, no report yet

Expected delivery times:

• Signal: about 15 minutes (typical)
• Scan: about 15-30 minutes (typical)
• Solutions: about 15-20 minutes (typical)

If delayed > 1 hour:

1. Check spam folder (hi@surmado.com may be filtered)
2. Check email address (verify order confirmation went to correct email)
3. Wait 2 hours total (load-dependent, some reports take longer)

If delayed > 2 hours: Contact hi@surmado.com with order details

What if something went wrong? We have automatic refund systems. If that doesn’t work, reach out to hi@surmado.com - we’ll make it right.

---

Issue 18: Charged for wrong tier (Essential vs Pro)

Symptom: Ordered Essential ($25) but charged for Pro ($50)

Fix: Contact hi@surmado.com immediately with:

• Order date/time
• Email address used
• Tier ordered vs tier charged

We’ll: Refund difference within 1-2 business days

---

Issue 19: Intelligence Token not included in report

Symptom: Received PDF report but no Intelligence Token

Fix:

1. Check email subject line (token is in subject: “Your Signal Report - SIG-2025-11-A1B2C”)
2. Check PDF page 1 (token usually listed at top or bottom)
3. Search PDF (Ctrl+F / Cmd+F for “SIG-” or “SCAN-” or “SOLUTIONS-”)

If still missing: Contact hi@surmado.com with order details

---

When to Contact Support

Contact hi@surmado.com if:

1. Delivery delayed > 2 hours
2. Billing issue (charged wrong amount, double-charged)
3. Intelligence Token missing from report
4. API returns 404 but token format is correct
5. Report contains errors (metrics don’t make sense, data seems wrong)
6. Refund request (within 7 days of order)
7. Technical integration help needed (Zapier, Make, n8n setup)

Include in support email:

• Intelligence Token (if you have it)
• Order date/time
• Email address used for order
• Specific issue description
• Screenshots (if helpful)

Expected response time: 1-2 business days (usually faster)

---

Self-Service Fixes (Try These First)

Before contacting support, try:

For Signal:

1. Read First Signal Report - Now What?
2. Check if Presence Rate 0% is normal (new/local business)
3. Verify Authority Score null is expected (if Presence Rate 0%)

For Scan:

1. Read First Scan Report - Now What?
2. Filter by high-severity issues only (don’t try to fix all 62)
3. Verify Overall Score < 70 is fixable (not broken)

For Solutions:

1. Read First Solutions Report - Now What?
2. Look for consensus (4+ models agreeing)
3. Focus on CFO/COO perspectives if others are unclear

For API:

1. Test token in browser: https://api.surmado.com/intelligence/{TOKEN}
2. Check token format (no spaces, correct pattern)
3. Read Intelligence Token JSON Reference

---

Common Misunderstandings

”My Presence Rate is 18% - does that mean I’m failing?”

No. 18% is baseline for many SMBs. It means AI mentions you in 18% of buyer queries. Your goal: improve to 25%, 35%, 45% over time. Not failure, starting point.

---

”Scan found 62 issues - does that mean my website is terrible?”

No. Most sites have 40-80 issues (if you test thoroughly). 5-15 are high-severity (fix those). 20-40 are low-severity (nice-to-have, not urgent). Don’t panic.

---

”Solutions gave me 3 different recommendations - which is right?”

All are valid. Recommendation #1 is highest consensus. #2 and #3 are alternatives (different risk/reward profiles). Choose based on your priorities (risk tolerance, budget, timeline).

---

”My competitor ranks #1 but has worse service - why?”

AI sees what’s published, not reality. If competitor has better schema markup, more content, clearer positioning on their website, AI will recommend them even if your service is objectively better. Fix what AI sees (your website).

---

”Intelligence Token API doesn’t need authentication - isn’t that insecure?”

Token IS the authentication. Anyone with token can access data (like a password). Don’t share tokens publicly. For internal use/automation, it’s fine.

---

Still stuck? Contact hi@surmado.com with your Intelligence Token and specific issue. We’ll help. If something went wrong with your report, we’ll make it right.