.d53b12e5.png&w=3840&q=75 "Trivas.ai")
by Trivas.aiHere's a scene that plays out at most DTC brands every Monday morning: someone opens a dashboard with 50 metrics on it, scans for three numbers, and closes the tab. The other 47 just sit there, generated by some integration someone set up eight months ago and never turned off.
It usually starts innocently. You connect Shopify, then Meta, then Google, then Amazon Seller Central. Someone builds a 40-tab spreadsheet or wires up a dashboard tool to pull it all together. Six months later nobody can tell you what half the tiles mean, let alone why they're there.
This post is for Shopify and Amazon sellers roughly in the $1M to $50M range, the same brands evaluating tools like Triple Whale, Northbeam, or Polar Analytics because their current setup has turned into exactly this kind of noise. If that's you, the fix isn't more dashboards. It's fewer metrics, chosen harder.
The test is simple: a dtc kpi dashboard metric earns its spot only if it changes a decision someone will actually make that week. If a number doesn't move a budget, a reorder, or a pricing call, it's decoration.
We'll walk through four categories that cover almost everything worth tracking: profitability, acquisition and paid media, retention and LTV, and operations and fulfillment. Everything else is optional at best.
Gross revenue and ROAS feel good to look at. They also hide unprofitable growth better than almost any other metrics on a typical dashboard.
The fix is contribution margin. Take revenue per order, subtract COGS, shipping, payment processing fees, and the variable ad spend tied to acquiring that order. What's left is what you actually get to keep before fixed overhead. A brand can be "profitable" on a P&L and still be bleeding cash on every new customer order if nobody's tracking this at the order level.
Blended CAC matters more than most people treat it. Not CAC by channel, blended across the whole business, because that's the number that tells you if your actual spend is sustainable against your actual margin. Too many teams optimize channel-level CAC while blended CAC quietly climbs.
MER, or marketing efficiency ratio (total revenue divided by total ad spend), is worth putting front and center too. It's harder to game than platform-reported ROAS because it doesn't rely on any single platform's attribution model. If MER is trending down while Meta's dashboard says ROAS is holding steady, trust MER.
The mistake we see constantly: brands track ROAS per channel without netting out discounts, returns, or Amazon referral and FBA fees. A campaign showing 4x ROAS on paper can be break-even or worse once you back out a 20% discount code and a 12% return rate on that SKU. If you want a cleaner read on this, the ROAS calculator is a quick way to sanity-check a channel number against what you're actually netting.
Blended CAC tells you if the business overall is healthy. It tells you nothing about where to move next month's budget. For that you need CAC by channel: Meta, Google, TikTok, Amazon Ads, each on its own line.
New customer ROAS and blended ROAS are not the same metric, and conflating them is one of the more expensive mistakes we see. Blended ROAS includes existing customers who were probably going to buy again anyway. New customer ROAS isolates the actual cost of growth. Brands that only watch blended ROAS tend to over-invest in retargeting, because retargeting makes blended ROAS look fantastic while doing very little to grow the customer base.
CPC and CPM trends deserve a spot on the dashboard too, even though they feel like upstream, less exciting numbers. They're early warning signals. CPMs creeping up for three weeks straight will show up in CAC eventually, but if you're only watching CAC you find out after the damage is done instead of while there's still time to adjust bids or creative.
One caveat that applies to every number in this section: platform-reported conversions are optimistic. Meta Ads Manager and Google Ads each take credit for conversions the other platform also claims, and both routinely overstate what a unified GA4 or warehouse-level view shows. If your acquisition metrics are pulled straight from each platform's own dashboard instead of a blended source of truth, treat every number with a grain of salt. This is exactly the kind of fragmentation a proper BI reporting layer is built to resolve, since it reconciles the channels against actual order data instead of trusting each platform's self-reported version of events.
Acquisition metrics tell you what a customer cost. Retention metrics tell you whether that cost was worth it.
Repeat purchase rate is the simplest version of this and it's criminally underused. If customers aren't coming back, your paid acquisition engine is working against an unlimited CAC ceiling with no payback. Pair it with 90-day and 180-day LTV cohorts, which show you whether the customers you acquired this quarter are actually paying back their acquisition cost on a realistic timeline, not just on day one.
For brands selling anything consumable or replenishable, subscription or reorder rate belongs on the dashboard too. It's one of the cleanest predictors of whether the unit economics hold up past the first sale.
LTV:CAC ratio gets thrown around as a benchmark, and 3:1 is the number everyone quotes. Treat it as a floor, not a target. A brand sitting right at 3:1 has very little room to absorb a bad quarter of rising CPMs or a shipping cost increase. If you're planning around LTV:CAC, aim meaningfully above 3:1, not just over the line.
Here's why most dashboards skip cohort-based LTV entirely: it requires joining order history across every channel a customer touched, not just whatever one platform's native reporting shows. Shopify's own analytics won't show you LTV for a customer who first converted through a Meta ad and later reordered through email. That join has to happen somewhere, and it usually doesn't happen in a spreadsheet past a certain point.
Profitability and acquisition get all the attention. Operations metrics quietly decide whether those numbers hold up.
Inventory turnover and sell-through rate matter a lot for any brand running Amazon FBA alongside a Shopify DTC store, because inventory sitting in two different systems makes it easy to overbuy in one channel while running lean in the other. Slow turnover ties up cash and, on Amazon specifically, starts racking up long-term storage fees that erode margin nobody's watching.
Return rate and refund rate by SKU deserve their own line, not just a blended number. A product with strong top-line sales and a 25% return rate can be a net drag on contribution margin even though the dashboard headline number looks great. Blended return rate across the catalog can hide a single bad SKU dragging everything down.
Fulfillment cost per order and shipping cost as a percentage of AOV are the quiet killers. A brand can hold gross margin steady while fulfillment costs creep from 8% of AOV to 14% over a year, and if that number isn't on the dashboard, it just shows up later as "margin somehow got worse" with no clear cause.
Stockout rate is worth tracking as a leading indicator, not an afterthought. A stockout on a top SKU shows up two weeks later as a CAC spike, because the acquisition engine is still spending against a product that isn't available to sell. If your acquisition metrics are swinging for no obvious reason, check inventory before you touch the ad accounts.
The filter is the same one from the intro, worth repeating here: does this metric change what someone does on Monday morning? If the answer is no, cut it. No exceptions for "it's nice to have."
Different roles need different dashboards, and trying to serve everyone with one view is part of why dashboards balloon to 50 tiles. A founder or CEO needs contribution margin, MER, and LTV:CAC, the numbers that answer "is this business healthy." A performance marketer needs CAC by channel, CPC and CPM trends, and channel-level ROAS, the numbers that answer "what do I do with next week's budget." Trying to cram both audiences onto one screen is how you end up with a dashboard nobody fully uses. (We cover this split in more detail for founders and CEOs specifically, since their reporting needs are consistently the ones getting buried under channel-level noise.)
Watch for vanity metrics sneaking onto the board deck version of the dashboard. Total impressions, follower growth, raw revenue with no margin context attached, these look impressive in a slide and drive exactly zero decisions. If a metric only exists to make a meeting go well, it's not a KPI, it's a talking point.
Most of this mess traces back to one root cause: data sitting in five different places. Shopify, Amazon Seller Central, Meta, Google, GA4, each with its own definitions and its own version of "conversion." Until that's reconciled somewhere, every dashboard is really just a manual reconciliation project with a nicer UI.
Spreadsheets are fine until they're not. They're flexible, cheap, and easy to customize for exactly one person's workflow. They also break down the moment you're pulling from more than two or three sources, and someone ends up spending three hours a week manually refreshing pivot tables that are stale again by Wednesday.
Native platform dashboards (Shopify's own analytics, Amazon Seller Central, Meta Ads Manager) are accurate for what they measure. The problem is they only measure their own channel. Shopify can't tell you blended CAC across Meta and Google. Amazon Seller Central has no idea what your Meta spend looked like last month. None of them were built to answer the questions that actually span the business, like true MER or new customer ROAS net of returns.
Purpose-built DTC analytics tools solve this by centralizing everything into a warehouse, so every channel is reporting against the same source of truth instead of five platforms each grading their own homework. If you're building this out for the first time, a custom dashboard setup that pulls Shopify, Amazon, and ad platform data into one place is usually the difference between a dashboard people trust and one they double-check against three other tabs before believing.
This is where Trivas fits. The core is a BI reporting layer built on Amazon Redshift, so contribution margin, blended CAC, and MER are all calculated from the same reconciled dataset instead of platform-reported numbers that disagree with each other. On top of that sits Wingman, the AI insights layer, which flags anomalies in these KPIs (a stockout about to hit, a CAC spike on a specific channel) instead of waiting for someone to notice a dip three weeks later while scrolling through tabs.
If you're rebuilding your dashboard from scratch, start with five: contribution margin, blended CAC, MER, LTV:CAC, and repeat purchase rate. That's it. That's the whole starter kit.
Adding metrics later is easy. Trimming them is what nobody ever does, which is exactly why most dashboards end up bloated in the first place. Start narrow, prove each metric earns its place by actually getting used, then expand.
If you want clear definitions for any of these before you build them out, the data dictionary has the formulas and edge cases spelled out so your team isn't arguing about what "repeat purchase rate" even means. And if you're past the spreadsheet stage and ready to stop reconciling five platforms by hand, that's worth a conversation sooner rather than later.
A dtc kpi dashboard's only job is to make the next decision obvious. Everything else on the screen is just decoration.
explore more insights
3 min read
3 min read
3 min read