Manifesto
The RateMyPrompts Manifesto
Everything starts with the prompts.
We love prompts, prompt design, prompt engineering and crafting something that initially appears abstract into something that has meaning and accurately reflects a business.
And there's a good reason for this!
Every GEO score, whether it's visibility or sentiment, depends on a set of prompts. It's the prompts that decide what gets measured and which competitors show up. Get them right and your GEO measurement gives you something meaningful. Get them wrong and, well, you're beautifully measuring the wrong thing. But how do you know which it is?
The existing GEO tools may take all the glory at the moment but it's the prompts that are the real unsung heroes of GEO.
The uncomfortable truth.
Unlike traditional search, GEO doesn't really have sufficient search or click data to show what people actually ask the LLMs. There's no agreed list of prompts and no reliable way of knowing whether a set reflects the questions that matter to the business. Different GEO tools, different GEO experts, even different LLMs will all give a very different set of prompts and therefore different sets of scores.
We know getting the right set of prompts is hard and that's really why we're here, to provide greater confidence that you're measuring the right thing.
What good prompts need.
Unlike our more SEO orientated brethren, we believe that a good prompt taxonomy starts with the business, not the data. It starts with its commercial priorities and decisions it wants to influence in AI. It then translates this into a series of structured prompts built around its customers, products & services, features and use cases. Then it uses the SEO data to fill the gaps.
And each prompt needs a clear purpose, because measuring visibility is a different job from measuring how AI understands your brand and measuring sentiment.
There is an art and science to prompt creation. If the prompt is too specific, it can lead the answer. But if it's too broad, there's a danger that it won't name any brand at all.
And what makes it harder is that GEO does require a decent volume of prompts to accommodate for different ways people search and the non deterministic nature of LLMs. But the more prompts you have, the more important the framework.
Generating 1,000 prompts is easy (thanks ChatGPT!) but creating 1,000 prompts that have balance, reflect all the commercially important areas of the business. And don't inevitably have some major gaps or skew the results is much, much harder.
Where this came from.
RateMyPrompts grew out of Zebora, a specialist GEO consultancy working with large brands and agencies including Jaguar Land Rover, Big Potato Games and Sykes Cottages. We spend weeks designing and refining prompt taxonomies for each client and developed an indepth scoring logic to check them. So we know first hand how difficult it is. And how autogenerated prompts almost always miss the mark.
To be honest, we didn't set out to build a prompt score engine. But we did set out to get GEO measurement right and so recognised the importance of the prompt taxonomy in this process.
What we believe.
Your GEO measurement is only as good as the prompts behind it.
And naturally, we've developed some pretty strong opinions about what that means.
- Prompts deserve a lot more attention than they're getting.
- The whole prompt set matters as much as each individual prompt.
- A prompt set should be explainable and stand up to scrutiny.
- Prompts need both volume and structure.
- Good prompt creation starts with the business and its commercial priorities.
Why RateMyPrompts exists.
We want this expertise to be available to any brand, agency or GEO specialist who wants to know whether their dashboard means what they think it means.
And we want to cut through some of the nonsense surrounding GEO measurement, giving people a straight answer on what's good, what's questionable and what needs fixing.
We're starting with visibility and discovery prompts, where most GEO measurement happens today.
RateMyPrompts.Give your prompts some love.