> ## Documentation Index > Fetch the complete documentation index at: https://docs.slingso.com/llms.txt > Use this file to discover all available pages before exploring further. # Your two scores > What Purchase Likelihood and Narrative Accuracy mean — and what to do when one is low. Slingso tracks two scores for every product, each from 0 to 100. Together they tell you not just *whether* you're winning the AI Shopping Channel, but *why*. Both update every week, each with a "vs last week" change so you can see which way you're moving. ## Purchase Likelihood — how likely is the purchase? **Purchase Likelihood** is your revenue signal: a 0–100 score for how likely your product is to be the one bought for a given shopper question — whether the shopper is a person deciding with AI's help, or an AI agent buying on their behalf. * A **high** score means you're very likely to win the sale — you're the product the shopper (or their AI) ends up buying. * A **low** score means the sale is going elsewhere — a competitor is the one being bought for that question. Slingso scores your competitors the same way, so you can see the gap — not just your number, but how far ahead or behind the product winning the sale. Purchase Likelihood is measured per shopper question, then rolled up per product and across your brand — so you can see exactly which questions you win and which you lose. ## Narrative Accuracy — is AI describing you right? **Narrative Accuracy** is your brand signal: a 0–100 score for how well the AI's description of your product matches the story you intend. * A **high** score means AI mentions the things that matter about your product, and gets them right. * A **low** score means AI is telling shoppers the wrong story — leaving out your key selling points, getting a detail wrong, or crediting one of your strengths to a competitor. "The story you intend" isn't a guess. It comes from **your [Brand Soul](/guide/brand-soul)** (how you want to be seen) and the key attributes on your product pages (the facts that should come through). On the [Products page](/guide/products) you'll see it as an **attributes-covered count** — how many of those key attributes AI actually gets across — plus exactly which ones it's missing or getting wrong. ## Reading them together The two scores point to different fixes. That's the whole value of having both: | Purchase Likelihood | Narrative Accuracy | What it's telling you | | ------------------- | ------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | Low | Low | AI doesn't really know your product yet. The fix is content. | | Low | High | AI describes you well, but the sale still goes to a competitor — usually price, availability, or where you're sold. Check the [channel & pricing view](/guide/products). | | High | Low | You're winning, but on a shaky story. Worth shoring up before it slips. | | High | High | You're winning for the right reasons. Defend it. | You don't have to work this out yourself — the agent reads the same signals and brings you specific opportunities. But seeing both scores tells you, at a glance, what kind of problem you're looking at. Your Brand Soul — the story the agent works from, and what Narrative Accuracy is measured against. How to tell when an approved fix actually moved your scores.