KnitKnot

# Benchmark products separately

How to add a product so it gets its own questions, competitors, and report section, and how per-product results roll up under your brand.

If your company sells more than one product, buyers ask AI about each one differently : different questions, different rivals, different winning claims. KnitKnot models this with [subjects](/docs/subjects/): your brand is one subject, and each product is its own subject underneath it. Adding a product gives it a dedicated question library, competitor set, and slice of the report, while everything still rolls up under the brand.

## Add a product

During [Benchmark setup](https://app.knitknot.ai/onboarding), review **Company & products** and include each product line you want measured. Confirm those choices before moving to **Competitors** and **Prompts**. After setup, review the product records under **Facts → Fact Ledger**.

Each included product needs relevant competitors and buyer prompts. Check the product filter in [Prompt library](https://app.knitknot.ai/prompts) before a run. If those inputs are missing, update the setup record before interpreting a product-level result.

## What "separate" means

Once a product exists, everything that hangs off it is scoped to it:

  • - **Questions** : the product has its own library; editing it doesn't touch your brand's questions.
  • - **Competitors** : its competitive field is tracked independently.
  • - **Scoring** : every response is scored against the product it belongs to, so a claim about one product never counts for or against another.

## How products roll up

A product recommendation is a brand win. When AI picks one of your products over a competitor : or picks one of your products over *another* of your products : that's a win for your company, not an internal loss. KnitKnot treats products that share a parent brand as one family for the purpose of who won.

In the report, this shows up as a company roll-up with a chip per product. Switch chips to drill into a single product's numbers; the brand itself is the roll-up, never its own chip. The [report guide](/docs/read-your-report/) covers reading the subject filter.

Next: [run a benchmark](/docs/run-and-schedule-benchmarks/) with the engines available to your workspace.

Docs navigation
Docs › Guides

Benchmark products separately

How to add a product so it gets its own questions, competitors, and report section, and how per-product results roll up under your brand.

Updated

If your company sells more than one product, buyers ask AI about each one differently : different questions, different rivals, different winning claims. KnitKnot models this with subjects: your brand is one subject, and each product is its own subject underneath it. Adding a product gives it a dedicated question library, competitor set, and slice of the report, while everything still rolls up under the brand.

Add a product

During Benchmark setup, review Company & products and include each product line you want measured. Confirm those choices before moving to Competitors and Prompts. After setup, review the product records under Facts → Fact Ledger.

Each included product needs relevant competitors and buyer prompts. Check the product filter in Prompt library before a run. If those inputs are missing, update the setup record before interpreting a product-level result.

What “separate” means

Once a product exists, everything that hangs off it is scoped to it:

  • Questions : the product has its own library; editing it doesn’t touch your brand’s questions.
  • Competitors : its competitive field is tracked independently.
  • Scoring : every response is scored against the product it belongs to, so a claim about one product never counts for or against another.

How products roll up

A product recommendation is a brand win. When AI picks one of your products over a competitor : or picks one of your products over another of your products : that’s a win for your company, not an internal loss. KnitKnot treats products that share a parent brand as one family for the purpose of who won.

In the report, this shows up as a company roll-up with a chip per product. Switch chips to drill into a single product’s numbers; the brand itself is the roll-up, never its own chip. The report guide covers reading the subject filter.

Next: run a benchmark with the engines available to your workspace.