If you’re new to Peec AI, we recommend reading “Setting up your prompts” and “Choosing the right prompts” first.
- A category prompt: This asks about a type of product (e.g., “best running shoes for beginners”. It measures every product you sell in that category at once.)
- A product prompt names one specific product (e.g., “Is the Nike Pegasus good for flat feet?”. It gives you a close-up on that one product.)
Step 1: Starting with categories
Every product in your catalog is tracked automatically. Peec AI matches each product to the category prompts it’s relevant to, so “best running shoes for beginners” measures every running shoe you sell, not just one.You don’t need a prompt for every product. Peec AI matches each product to the category prompts it’s relevant to, so “best running shoes for beginners” measures every running shoe you sell, not just one.This is why you don’t need a prompt for every product. A single-category prompt already covers all products in that category.
- Every product is measured through its category’s prompts regardless of how they’re worded.
- A clean category structure mirrors how you already organize your range, so the data stays readable. Build the set around the intents you want to track, with categories as the frame.
Step 2: Build each category’s core set
For each category, start with a small core of reliable prompts. Lead with non-branded discovery, then add a smaller branded slice. Non-branded discovery (the priority): These carry most of the day-to-day value, because they test whether AI recommends you when the shopper hasn’t named a brand:- “What’s the best [product category] for [key use case]?” The broad recommendation question category leaders should win. A few variations on the use cases and constraints that matter most in that category (see Step 3).
Peec auto-classifies them under the branded tag, so you can filter to non-branded any time to read your true discovery numbers, no separate project needed.
- “What are the best alternatives to [your brand]?” Captures shoppers who already know you and are comparing options.
- One or two comparisons against your top competitors, phrased as real buying decisions: “I’m training for my first marathon, should I choose [your product] or [competitor product]?” Comparisons often produce a side-by-side table, which is a good source of attribute data.
- Small catalogs (up to about 100 products): around 3 prompts per product is workable, folded into the category structure.
- Hero products: Focus product prompts on your highest-priority products, with comparisons and purchase questions.
- The long tail: Track a handful of product prompts for non-hero products too, grouped under a single tag, so you can see how they perform without crowding the set.
Step 3: Think in dimensions
Once the core is in place, expand coverage across three dimensions:| Dimension | How many | What it looks like |
|---|---|---|
| Category | Your top 3 to 5 to start (up to 10 to 30 for large catalogs) | One topic per category, mirroring your catalog’s tree. If your catalog is one big category, use product lines (e.g. “iPhone 17”, “Air Jordan 1”) as the topics instead. |
| Buyer context | 2 to 4 per category | The constraint that changes the answer. Keep them within one category so the set stays focused, e.g. for running shoes: “for beginners”, “for flat feet”, “for trail running”, “under €120”. |
| Intent stage | 2 to 3 per buyer context | How close the shopper is to buying, from weighing options (“best running shoes for flat feet”) to ready to purchase (“where to buy the Nike Pegasus in Germany”). Peec AI classifies intent automatically; keep the set mostly commercial and transactional. |
- Prioritize distinct buying situations: A new intent or context adds more than a reworded duplicate. That said, tracking a few variants of your most important prompts is worthwhile: it stabilizes the day-to-day numbers, since a single prompt on a single day is noisy, whereas a handful read steadily.
- Name something shoppable in every prompt: A product category or a specific product, not just a vague need (e.g., “Best Pilates socks for beginners” triggers shopping results, while “what do I need for my first Pilates class?” usually doesn’t).
- Prompts with commercial intent: Shopping recommendations appear on more than half of commercial prompts, versus roughly one in ten informational ones. Keep some informational coverage, but focus the set on discovery and purchase decisions.
Not every prompt will generate a shopping recommendation, and that’s normal. Shopping results are most likely for clear product-focused queries, especially physical goods rather than services.