Stop overpaying for Shopify analytics

Uninstalling deletes everything we hold about your store: orders, products, costs, the lot. Not archived, not anonymised. Gone.

It depends on your store size. Under 10,000 orders you're usually looking at a couple of minutes. A large store with years of history can take a few hours, and you don't have to wait for it, the dashboard fills in as the data lands.

Ad spend: Meta, Google, TikTok, Snapchat, Pinterest, Microsoft and AppLovin.

Shipping costs: ShipHero, or a CSV of your own carrier rates.

Affiliates: GoAffPro. Email: Klaviyo.

We pull COGS directly from your Shopify store. You can also set custom COGS between specific periods, so you're not limited by Shopify's one-COGS-to-rule-them-all.

You can also define COGS for deleted products and variants, ensuring your historical data remains accurate.

Additionally, you can set a fallback COGS percentage to apply when you haven't input a specific COGS.

You can input an average shipping cost per order, or input your full cost matrix from your carrier by defining the cost per weight interval. Kleio then calculates the actual shipping cost based on the weights of each order.

You can also integrate directly with ShipHero to pull your shipping costs automatically.

Kleio has the most flexible variable costs in the game. You can define variable costs per order, per line item, per refund, per refunded line item, as a percentage of sales, as a percentage of COGS, by order weight (in intervals), or by line item quantity (in intervals).

You can also set start and end dates for each variable cost, so you're not locked into the same cost forever.

Variable costs can also be applied to specific countries, products, shipping methods, and more.

Yes! You're always more than welcome to reach out to me directly at help@getkleio.com.

Attribution pretends to be objective. It isn't.

Every attribution model is subjective. It's an opinion baked into code by the developer who built it. They decide what to allocate where. Change those rules, and the story changes. That's storytelling, not truth.

We also don't believe in separating optimization from reporting. Your ad platforms optimize using their own data and models, regardless of what an external attribution tool says. So you end up reporting on one version of reality… while optimization happens in another.

Meta is extremely good at what they do. They have more data on this than anyone else on the planet, and that's exactly why their targeting and optimization algorithms perform so well for advertisers. There's no reason to believe a third party attribution tool, with less data and no idea about how Meta's algorithms work, can consistently provide better attribution than the platform itself.

There's also an incentive problem. Attribution software has to justify its price, and the easiest way to do that is to look meaningfully different from the numbers you get for free inside ad platforms. We're not saying anyone is acting in bad faith, but as the saying goes: show me the incentive, and I'll show you the outcome.

We've seen this too many times to ignore: ads get turned off completely, yet attribution tools keep assigning them hundreds of conversions more than a month later. That doesn't sound right. Here's an example of that.

So, instead of building attribution just so we can charge hundreds of dollars a month, we focus on what actually moves the needle, and give it to you at a fair (cheap) price.

Kleio ships an MCP server that connects your AI assistant (Claude, Cursor, Windsurf, or anything that supports the MCP protocol) directly to your store data. Instead of digging through dashboards, you ask questions in plain English: "What products should I push this month?", "How's my CAC trending?", "Which SKUs are losing money?"

Your AI pulls P&L breakdowns, product metrics, ad spend analysis, and more straight from Kleio. It's like having an analyst on call 24/7.

"I plugged Kleio into Claude. Every two weeks it pulls my full P&L, product margins, cohort health, and ad ROAS — and synthesizes it into strategic recommendations. In minutes." — Chris on X ↗

Read more in the MCP server docs.

Yes. They solve different problems.

When you want to check your margins before a meeting, you don't want to type a prompt and wait for a response. You open the dashboard, see your numbers, and move on with your day.

But when you need to dig deeper, LLMs don't get tired. They don't miss connections. And they handle way more data than you ever will. They just synthesize.

The dashboard is your daily driver. The MCP server is your analyst on call.

We don't have a native app, but the Kleio dashboard is fully responsive and works great on mobile. Open app.getkleio.com in your mobile browser, tap Share → Add to Home Screen, and it works just like an app.

Read more in the mobile app docs.