Welcome to Ranketta
How Ranketta collects data
AI responses are not consistent. Ask the same prompt twice and you'll often get a different answer. That's the nature of LLMs. We account for this in two ways:
Percentages, not exact counts. We express all metrics as percentages (visibility, share of voice, sentiment score) because they suit the probabilistic nature of LLMs better than fixed numbers. A "25% visibility" is meaningful; a single "yes" or "no" from one query is not.
Frequent queries, not single snapshots. We run your prompts across AI platforms like ChatGPT, Gemini, and Copilot very frequently, not just once. The more responses we collect, the more accurate and stable your metrics become.
These prompts are conversational questions like "What's the best CRM for marketing agencies under 50 people?" instead of keywords like "CRM software." This matches how people actually talk to AI tools.
Unlike traditional analytics tools that rely on APIs, Ranketta uses advanced UI scraping technology to interact with AI models exactly as real users do. This approach provides several key advantages:
- Authentic user experience: We interact with AI models through their web interfaces, capturing the same responses users see.
- Real-world accuracy: Our data reflects actual user experiences, not sanitized API responses.
- Comprehensive coverage: We can access models that don't offer public APIs or have limited API functionality.
Our UI scraping technology simulates real user interactions, ensuring the data we collect matches what the average user sees when they use these AI tools.
For most AI platforms, this means that they decide which model to use and if they perform a web search or not. This is the reality for a logged-out user and the approach that's closest to the average user's experience.
Why we use the web interface, not APIs
Many tools pull data from official APIs. The problem: what the API returns often isn't what users see. Providers may serve different answers, different citations, or different formatting through their API than in the actual chat. Your metrics would reflect a backend channel, not the real experience.
We use the same interface your customers use: the web. We send prompts through the browser and capture the responses that appear on screen. That means:
- Same journey as your users. Same input, same output. No separate API pipeline that might behave differently.
- Works where APIs don't. Some platforms have no public API or heavily restrict it. We still measure them.
- Not tied to API changes. Providers update or deprecate APIs all the time. Our approach keeps working.
Your visibility data comes from real user flows, not from a filtered API feed.
FAQ
Why can't I get the same answer when I ask ChatGPT myself?
AI answers aren't deterministic. The same prompt can return a different response depending on the session, personalization, and whether the model runs a live web search. Ranketta runs each prompt many times from a logged-out, average-user perspective, so instead of one lookup you get a stable percentage that reflects the real spread of answers.
Does Ranketta use official APIs from OpenAI, Google, and other providers?
No. APIs can return different answers, different citations, and different formatting than the actual chat interface, so measuring through an API wouldn't reflect what your customers actually see. Ranketta scrapes the same web interface real users interact with, capturing the responses as they appear on screen.
How often does Ranketta query my prompts?
Very frequently, not as a single snapshot. We run your prompts across AI platforms continuously, and the more responses we collect, the more accurate and stable your visibility metrics become.