Zero Mentions Despite AI Presence
This is the most frequently reported issue. You might say “I know AI mentions my brand, but Trakkr shows zero mentions.” Understanding why this happens helps resolve the problem quickly.Why This Occurs
Prompt Specificity Mismatch Your prompts may not match the contexts where AI models actually mention your brand. For example, if you’re a project management tool but your prompts ask about “business software” generally, AI models might mention you in response to “team collaboration tools” instead. Brand Name Variations AI models might know your brand by different names or variations. They could mention “Apple Inc.” while you’re tracking “Apple,” or refer to “McDonald’s Corporation” when you’re monitoring “McDonald’s.” Similarly, they might use abbreviations, acronyms, or informal names. Market Positioning Differences Your prompts might target a broader or narrower market than where your brand actually has recognition. If you’re a specialized B2B tool but your prompts ask about consumer software, AI models won’t make the connection. Training Data Limitations AI models have training data cutoffs, and newer brands or recent rebranding efforts might not be well-represented. Additionally, some models may have stronger knowledge in certain industries or geographic regions.Diagnostic Steps
Start by testing your brand recognition directly. Try asking the AI models simple questions like “What is [YourBrand]?” or “Tell me about [YourBrand]” to see if they know your brand at all. If they don’t recognize your brand name, the issue is fundamental brand awareness in AI training data. Next, examine your prompts for relevance. Look at each prompt and honestly assess whether your brand should appear in the top 10 responses. If you’re a niche player in a competitive market, you might need more specific prompts that better reflect your actual market position. Test alternative brand names by checking if AI models know you by variations, abbreviations, or your full legal name. Sometimes brands are better known by their parent company name or a specific product name.Solutions
Refine Your Prompts Make your prompts more specific to your actual market segment. Instead of “best business software,” try “best project management tools for remote teams” if that’s your specific niche. Use the language your customers actually use when searching for solutions like yours. Test Multiple Contexts Create prompts covering different aspects of your business. If you’re mentioned in response to questions about specific features, use cases, or industry applications, make sure you have prompts covering those areas. Check Name Recognition If AI models know you by a different name, consider whether you should update your brand name in Trakkr or create content that reinforces your preferred brand name usage. Start Broader, Then Narrow Begin with prompts where you’re confident you should appear, even if they’re very specific to your niche. Once you see mentions, gradually expand to broader industry questions.Prompt Management Issues
Prompts Not Running or Infinite Spinners
You might encounter issues with prompts getting stuck or not executing properly. This typically happens during the prompt creation or editing process. During Setup If your prompts seem stuck during initial setup, wait for the full setup process to complete. The “Running your first report” step can take up to 30 seconds, and interrupting it may cause issues. Adding New Prompts When you add prompts after initial setup, they won’t run immediately. New prompts are included in the next daily report, which means you’ll see results the following day. There’s no way to trigger immediate execution. Editing Existing Prompts If you’re having trouble saving edits to existing prompts, ensure you’re not exceeding character limits or using special characters that might cause formatting issues. Save changes one at a time rather than making multiple edits simultaneously.Site Addition Problems
You might experience issues adding your website during setup or when creating new brands. This usually relates to website accessibility or format issues. Website Format Ensure you’re entering the complete URL including “https://” or “http://”. Enter “https://yoursite.com” rather than just “yoursite.com.” Website Accessibility Your website needs to be publicly accessible for Trakkr to analyze it. Sites behind login walls, under construction pages, or with strict bot blocking may cause issues. Temporary Issues If your website is temporarily down or experiencing issues during setup, this can cause the analysis to fail. Try the setup process again when your site is fully accessible.Scores Don’t Match Direct AI Queries
Why Trakkr results differ from manual testing If you ask ChatGPT or Claude the same prompt directly and see different results than what Trakkr shows, this is completely normal. Here’s why: Web search enabled (Free vs Paid plans) On the Free plan, Trakkr doesn’t use web search enabled prompts. However, when you test manually on ChatGPT, web search is often enabled by default. This means ChatGPT may show brands more frequently when you test manually, while Trakkr Free results won’t include these web search-enabled responses. Paid plans have web search enabled for models like ChatGPT and collect citations from those searches—this is why paid plan results more closely match what you see when testing manually. Learn more about citations → Different query timing Trakkr runs your prompts at specific times each day (typically overnight). When you test manually, you’re querying at a different time, and AI models can have slight variations in their responses between sessions. Model version differences AI models update frequently, and different API endpoints may use slightly different model versions. Trakkr uses the same models consistently, but manual queries might hit different versions. Response randomness AI models have inherent randomness in their responses. The same prompt asked multiple times can produce slightly different rankings, especially for brands in similar positions. Context and conversation history When you ask an AI model directly, your conversation history and context can influence responses. Trakkr queries are fresh, isolated requests without prior context. What to do Focus on trends over time rather than comparing individual queries. Trakkr’s value is in tracking consistent patterns and changes over days and weeks, not matching a single manual query exactly. If you consistently see major discrepancies over multiple days, check that your prompts in Trakkr exactly match what you’re testing manually.Score Fluctuations and Anomalies
Understanding Normal Variation
Some score fluctuation is completely normal and expected. Different AI models update their training data at different times, so your scores may vary as models incorporate new information. Additionally, AI models can have slight randomness in their responses, meaning the same prompt might generate slightly different results on different days. Market dynamics also affect scores. New competitors entering the market, existing competitors launching major campaigns, or industry events can influence how AI models rank brands.When to Be Concerned
Sudden Dramatic Drops If your visibility or presence scores drop by more than 20 points overnight without obvious cause, investigate which specific prompts or models are affected. This might indicate a change in AI model behavior or new competitive dynamics. Consistent Decline Over Weeks A steady decline over multiple weeks suggests a real trend rather than random variation. Analyze whether this coincides with competitive launches, market changes, or changes to your own prompts. Extreme Scores Visibility scores over 100 or presence scores over 100% indicate a calculation error and should be reported to support. Similarly, identical scores across multiple days might suggest a data processing issue.Troubleshooting Score Issues
Review your prompt performance individually to identify which specific prompts are causing unusual results. Sometimes one problematic prompt can significantly affect your overall scores. Check the Query View for affected prompts to see exactly what AI models are returning. This helps you identify whether the issue is with your brand mentions specifically or with the AI models’ responses generally. Compare your performance across different AI models. If only one model shows unusual results, this might indicate model-specific issues rather than problems with your brand or prompts.Technical Platform Issues
Dashboard Loading Problems
If your dashboard isn’t loading properly, first check your internet connection and try refreshing the page. Browser cache issues can sometimes cause display problems, so try clearing your browser cache or opening Trakkr in an incognito/private window. Some browser extensions, particularly ad blockers or privacy tools, can interfere with dashboard functionality. Try disabling extensions temporarily to see if this resolves the issue.Report Generation Failures
Daily reports occasionally fail due to temporary API issues with AI models or high system load. If you notice a report didn’t generate on a particular day, it should resume automatically the next day. Multiple consecutive days of missing reports indicate a more serious issue that requires support intervention.Data Timing and Freshness
When is data collected? Trakkr runs daily reports automatically, typically overnight. Your data updates once per day, not in real-time. This means changes you make to your website, content, or prompts won’t appear immediately—they’ll be reflected in the next day’s report. Why isn’t my data real-time? AI model queries are computationally expensive and time-consuming. Running real-time queries for all users would be impractical and costly. Daily reports ensure consistent, reliable data collection while keeping costs manageable. How long until changes appear?- New prompts: Included in the next daily report (typically the following day)
- Content changes: May take several days to weeks to appear, as AI models need to crawl and index your updated content
- Technical fixes: Similar to content changes, AI models need time to re-crawl and process your site
- Prompt edits: Changes take effect in the next daily report
- Crawl your updated content
- Process and index the changes
- Incorporate new information into their responses
Account and Setup Issues
Multiple Brand Setup
While Trakkr supports multiple brands per account, each brand requires separate setup through the same process. You might expect to add multiple brands simultaneously, but each brand needs individual configuration with its own prompts and analysis.Adding Competitors
Why you can’t directly add competitors Trakkr doesn’t track brands directly—it tracks prompts. Brands appear as an output of the system, not an input. When you create prompts about your industry, Trakkr analyzes AI model responses to see which brands get mentioned and how they’re ranked. This means there’s no way to manually “add a competitor” to your tracking. How competitors appear Competitors automatically appear in your data when AI models mention them in response to your prompts. If you want to ensure specific competitors show up in your analysis, you need to use prompts that are likely to surface those brands. For example, if you’re tracking “best project management tools” and a competitor is well-known in that space, they’ll naturally appear in the results. Alternative: Using multiple brands for competitive tracking Larger brands often use the Scale plan, which includes 10 brand slots, to track competitors strategically. Here’s how it works:- Use your first brand slot for your actual brand (e.g., “Nike”)
- Use the remaining 9 slots to set up your main competitors as separate brands (e.g., “Adidas”, “Reebok”, “Puma”, etc.)
- Each competitor brand gets its own set of prompts, allowing you to see the world through their eyes
- This gives you comprehensive competitive intelligence by tracking how AI models view each competitor independently
Plan Limits and Feature Restrictions
Why can’t I add more prompts? Each plan has prompt limits per brand: Free (5 prompts), Growth (50 prompts), and Scale (50 prompts). These limits are designed to ensure comprehensive but focused tracking. If you’ve hit your limit, review your existing prompts and remove or consolidate ones that aren’t performing well. You can modify prompts at any time, so you’re not locked into your initial set. Why can’t I add more brands? Free and Growth plans include 1 brand. Scale includes 10 brands. Additional brands can be added for $39/month each on Growth and Scale plans. If you need more brands, consider upgrading to Scale (which includes 10 brands) or adding individual brands to your current plan. Why isn’t this feature available? Some features are only available on paid plans. For example:- Citations, sentiment tracking, and crawler analytics require Growth or Scale
- API access is exclusive to Scale
- Team features require Scale
Prevention Tips
Setting Realistic Expectations
Remember that AI brand tracking reflects how AI models currently understand your brand based on their training data. If your brand isn’t well-known online or in your industry, this will be reflected in your scores. Your scores naturally fluctuate as AI models update and as market dynamics change. Focus on long-term trends rather than daily variations, and remember that improving AI visibility often takes time and consistent effort.Best Practices for Stable Tracking
Start with prompts where you’re confident your brand should appear, then gradually expand to more competitive or broader queries. Monitor your prompt performance regularly and refine those that consistently underperform. Keep detailed notes about changes you make to prompts so you can correlate score changes with specific modifications. This helps you understand what works and what doesn’t for your brand. Maintain consistent tracking by avoiding frequent major changes to your prompt set. While optimization is important, constantly changing prompts makes it difficult for you to identify meaningful trends in your data.Understanding Your Scores
Learn more about how scores are calculated and what they mean
Optimizing Your Prompts
Improve your prompt strategy to get better tracking results