What makes a prompt effective?

A prompt is just the message you send a language model. A good prompt is one that gets you what you actually wanted on the first or second try. The single biggest factor is being specific.

Anthropic’s guidance puts it plainly: be clear and direct, and being specific about your desired output helps the result. If you want behavior that goes above and beyond, ask for it rather than expecting the model to infer it from a vague prompt.^[1] OpenAI’s guide makes the same point for its models, recommending precise, explicit instructions over loose ones.^[2]

A useful mental model from Anthropic: think of the model as a brilliant but brand-new employee who does not yet know your norms or your workflow. The more precisely you explain what you want, the better the result.^[1] A new hire who is told “make a chart” will make something, but probably not the chart you had in mind.

So the difference between “Create an analytics dashboard” and “Create an analytics dashboard. Include as many relevant features and interactions as possible. Go beyond the basics to create a fully-featured implementation” is not politeness or length for its own sake. The second one tells the model what “done” looks like.^[1]

Specific is not the same as long, which is easy to conflate. A three-paragraph prompt that never states the audience, the length, or the format is still vague, just at greater length; a single added sentence naming those three things does more work than doubling the surrounding prose. The test is not how many words came before the request, it is whether a reader could point to the exact spot that answers “how would I know if this was done right.”