How to Know If Your DTC Business Is Healthy, Not Just Growing

DTC has a metric imbalance problem.

The metrics that get the most attention are the ones that are easiest to see. ROAS, CAC, and conversion rate update in real time, live on polished dashboards, and tie directly to decisions that marketing teams make every day. They're non-negotiable, especially early. If you can't acquire customers at a viable cost, nothing else matters.

But they only answer one question: is acquisition working?

They don't tell you whether the orders you're shipping are profitable. They don't tell you whether return rates are quietly compressing your margin. They don't tell you whether the inventory you bought last quarter is going to force a markdown cycle. And they don't tell you whether the customers you're acquiring are actually coming back.

The metrics that answer those questions sit in different systems, updated on different schedules, and rarely connected to each other. Most DTC operators know they should be watching them. Most don't have a clean way to do it.

This is the framework. Not a list of every metric worth tracking, but a structure for understanding which metrics matter, in what order, and what they tell you about the health of your business, not just your marketing.

One commitment before we get into it. Every metric here gets explained in plain English first. The formal definitions matter and I'll give them, but if you came up through marketing rather than finance, that shouldn't be a barrier to understanding what you need to watch.

The core idea: three tiers, two directions

Think of your metrics as three layers.

At the top are Outcomes. These are the numbers that tell you if the business is actually healthy. Not if marketing is working. If the business is healthy.

In the middle are Drivers. These explain why the Outcomes are moving. When contribution margin drops, you look here to find out which lever moved.

At the base are Signals. These are leading indicators. They move before the Drivers shift, and long before the Outcomes deteriorate. They're your early warning system.

Here's the most useful thing to know about how the three layers work together.

Read top-down to diagnose. If a Tier 1 metric starts falling, you go into Tier 2 to find out which lever moved. If you can't see it there, you check Tier 3 for what may have been signalling the problem weeks earlier.

Read bottom-up to predict. A Tier 3 Signal starts moving. CPM is creeping up. Return request rate is ticking. Email engagement is dropping. Nothing has shown up in your headline numbers yet. But it's coming, and you have a window to act.

Most operators spend almost all their time in Tier 2. They manage CAC, ROAS, conversion rate. That's not wrong. But it means they're usually diagnosing problems after the damage is done, and never seeing them coming in advance.

Tier 1: Outcomes

The scorecard. These tell you if the business is fundamentally healthy.

These are the numbers your board, investors, or bank will look at. More importantly, they're the numbers that tell you if what you're building is actually sustainable, regardless of what the marketing dashboard says.

Net Revenue

In plain English: the money that actually lands in the business. Not the gross order value before returns. Not GMV. What stays.

Formally: total revenue after discounts and returns are deducted. This is your starting point for every downstream calculation. Contribution margin, LTV, and revenue per session all depend on getting this number right. Brands that use GMV as their primary revenue number routinely overstate margin and understate the real cost of returns.

Gross Margin

In plain English: after you pay for the product itself, what's left? This is the ceiling for everything else in your business.

Formally: revenue minus cost of goods sold, expressed as a percentage.

For most physical goods DTC, 50%+ is a working threshold. Below that, covering acquisition costs, fulfilment, and returns while staying profitable becomes very hard. A high-AOV, low-return category can operate at a lower gross margin than a high-return fashion brand, but that requires honest modelling. If you're below 45%, you need a clear explanation for why your model is the exception.

Contribution Margin

In plain English: after you fulfil the order, process any returns, and pay for everything that touched that transaction, did you actually make money?

Formally: gross profit minus all variable costs per order. Two things matter here that most P&Ls get wrong. First, COGS means landed cost (product plus inbound freight, duties, and tariffs), not the catalog cost most brands type into Shopify. Using catalog cost typically overstates CM1 by 5 to 10 points, more in categories hit by recent tariff changes. Second, the variable cost stack includes fulfilment, 3PL pick-and-pack and storage, packaging, payment fees (typically 2.9% of revenue plus a fixed per-transaction fee), last-mile surcharges, returns processing and reverse logistics, and any affiliate or influencer commissions tied to the transaction. Order-level discounts are already netted in Net Revenue and should not be subtracted again.Two layers worth knowing.

CM1 is the margin on each order before marketing spend is accounted for.

CM2 takes CM1 and subtracts total marketing spend: paid media, affiliate, influencer, and any marketing cost that scales with growth. This is the number that tells you whether growth is actually profitable, at a blended level. Per-channel or per-campaign CM2 is fragile because attribution noise and the judgment call of variable-versus-fixed marketing compound into numbers that look precise but are mostly opinion. Treat CM2 as a blended outcome, not a campaign scorecard.

Most operators calculate CM1 and call it contribution margin. The question worth asking is whether CM2 is positive.

Why this matters more than most operators realise: a brand with 62% gross margin, a 28% return rate, a 2.5% payment fee, and a heavy affiliate channel can easily land at sub-20% contribution margin. A lot of growth-first brands found this out after two years of apparently strong revenue.

A useful check: take your last three months of net revenue, subtract every cost that varies directly with order volume, and see what remains. That number is your real per-order margin.

EBITDA

In plain English: if a bank, investor, or potential acquirer wants to know whether your business model actually works, this is the number they look at. It strips out tax, debt interest, and depreciation to show the underlying operating performance.

Revenue growth with declining EBITDA is a red flag, not a growth story. This becomes critical earlier than most operators expect. From around the 10 to 20M revenue mark, when headcount and infrastructure start to scale with the business, this is what separates a business that is building something from one that is spending its way to growth.

Free Cash Flow

In plain English: the clock. Every other metric is theoretical if cash runs out.

You can have perfectly healthy unit economics and still run out of cash if the timing of inventory payments and revenue collection does not align. Scenario-plan your runway constantly. A business with three months of cash and strong contribution margin is not the same as one with 18 months of runway.

Net Promoter Score (NPS), stage-dependent

In plain English: how likely are your customers to recommend you? A Tier 1 health metric at scale, but only if you're actually running a methodologically sound NPS program.

NPS predicts repeat purchase rate, word-of-mouth acquisition, and long-term customer value. A declining NPS shows up in CAC and repeat rate months before it shows up in revenue. The caveat: below roughly $20M in revenue, most brands don't run a real NPS program, and a patchy one is worse than none. For brands in the 1M to 20M range, use Repeat Purchase Rate, Return Rate, and Tickets per Order as the leading proxy. From $20M and up, when you can run a proper program, treat NPS as a board-level outcome alongside EBITDA."

Tier 2: Drivers

The levers. When an Outcome moves, you look here to find out why.

These are the metrics you actively manage. They're split into five categories because each one explains a different part of what drives your Tier 1 outcomes. In many DTC businesses, Acquisition and Retention both sit within a marketing function, but they answer different questions. The diagnostic value comes from keeping them separate.

Acquisition

CAC (Blended): total marketing spend divided by new customers acquired. Blended, not by channel, because channels influence each other in ways that single-channel attribution misses.

CAC Payback Period: how many months until you recover the cost of acquiring a customer? Under six months is healthy for most DTC categories. Over 12 months, in any category, is a cash risk that compounds as you scale.

LTGP:CAC Ratio: the long-term view. Lifetime Gross Profit divided by customer acquisition cost. LTGP is cleaner than LTV because it already accounts for gross margin. Two brands with identical LTV:CAC can have radically different economics if one runs at 30% gross margin and the other at 70%. Under 3:1 LTGP:CAC is typically a warning sign for non-subscription DTC. Under 2:1 means the economics may not be viable at scale.

Retention

Repeat Purchase Rate: the single strongest indicator of whether your product delivers on its promise. A flat or declining repeat rate is a product signal, not a CRM problem. No amount of email optimisation compensates for a product customers don't want to come back for.

Email/SMS Revenue Contribution: what percentage of total revenue comes through owned channels? For physical goods DTC, email alone should drive 20 to 35% of revenue. Below 15% means you're renting your audience from paid platforms, and structural vulnerability to rising CAC follows.

Revenue per Subscriber: revenue generated per email subscriber over a rolling period. Declining revenue per subscriber is an early signal of list fatigue, weakening product-market fit, or a CRM programme that hasn't kept pace with list growth.

Customer Reactivation Rate: the proportion of lapsed customers successfully reactivated each period. A declining reactivation rate means your CRM is losing customers it once could recover, and LTV:CAC will follow.

Commerce

AOV (Average Order Value): higher AOV directly improves contribution margin per order. Bundling and upsells are structural margin plays, not just revenue tactics.

Conversion Rate: significant swings signal product-market fit issues, pricing problems, or UX friction. Also the metric most sensitive to paid traffic quality.

Cart Abandonment Rate: the percentage of sessions that add to cart but don't complete purchase. Directly actionable through abandoned cart email and SMS flows, which typically recover 5 to 15% of abandoned sessions at near-zero marginal cost. A rising abandonment rate signals checkout friction, payment issues, or pricing sensitivity, and each has a different fix.

Net Revenue per Order: gross revenue minus returns and refunds, divided by orders fulfilled. Your real top-line number at order level.

Operations

Return rate can make or break contribution margin, particularly in fashion, apparel, and homewares. A 30% return rate doesn't reduce revenue by 30%. It doubles fulfilment costs on those units and compresses the margin on every other order in the same batch.

Return Rate: the percentage of completed returns as a proportion of orders fulfilled. This is the lagging measure. Track it alongside return processing cost per unit because the margin impact is always higher than the headline rate suggests.

Note: Return Rate in Tier 2 is the damage already done. Its leading counterpart, Refund and Returns Request Rate, sits in Tier 3 and typically moves three to six weeks ahead. More on that below.

Fulfilment Cost per Order: what each order actually costs to pick, pack, ship, and handle. This should improve as you scale through volume leverage. If it isn't declining as revenue grows, operational efficiency isn't keeping pace.

On-Time Delivery Rate: the percentage of orders that arrive within the committed delivery window. Directly upstream of NPS, post-purchase CSAT, and repeat purchase rate. A late delivery is often the only post-purchase interaction a customer has with your brand. Best-in-class DTC operates at 95%+. Below 90%, the impact will show up in CSAT and reactivation rate within six to eight weeks.

Inventory

Inventory Turnover: how many times you sell through your inventory in a year. Low turnover ties up cash and increases the risk of having to discount to clear stock.

Sell-Through Rate: the percentage of inventory sold at full price. Everything else gets discounted or written off, and both outcomes compress contribution margin.

Days Inventory Outstanding: how long stock sits before it sells. Think of it as the cash cost of slow-moving inventory. The longer stock sits, the more capital it ties up with no return.

Stockout and Overstock Rate: two sides of the same planning failure. Stockouts kill conversion silently: the session happened, the sale didn't, and it's invisible in your traffic data. Overstock forces markdowns and trains customers to wait for promotions. Track both by SKU. Stockout above 5% on core SKUs is a silent revenue leak. Overstock above 20% of inventory value almost always means a markdown cycle is coming.

Tier 3: Signals

The early warning system. These move first.

This is the tier most operators ignore. Not because it's unimportant, but because the data is harder to see. It's often spread across platforms that don't connect to each other.

That's also exactly why it matters. By the time a problem shows up in your Tier 1 Outcomes, you've already lost months of response time. These metrics exist to give you that time back.

Rising CPM and CPC Trends: before CAC deteriorates, media costs move. Watch the trend week-on-week, not the snapshot.

Blended ROAS Trend: a deteriorating ROAS trend, channel by channel, is an early warning that blended CAC is about to rise. As a snapshot, ROAS is misleading: a 4x ROAS on a low-margin product is not the same as a 4x ROAS on a high-margin one. As a directional trend, it typically moves two to four weeks ahead of CAC. That gives you time to reallocate spend before the damage is done.

New vs Returning Customer Ratio: a growing proportion of repeat buyers is healthy. A shrinking proportion means you're burning acquisition budget to replace customers you're not retaining.

Add-to-Cart Rate: signals demand before it translates to revenue. A sustained drop is an early warning on conversion, often caused by traffic quality rather than the product itself.

Email Engagement Trends: declining open and click rate trends predict a drop in owned-channel revenue four to six weeks before it appears. Most operators treat open rate as a vanity metric. As a trend, it's one of the clearest signals that your retention engine is losing effectiveness.

Refund and Returns Request Rate: this is the leading counterpart to Return Rate in Tier 2. The distinction matters. The request rate measures the problem as it is forming. Return Rate measures the damage after it has happened. Problems show up here three to six weeks before they appear in return rate or net revenue. If this number spikes, your contribution margin for that cohort is already compromised.

Tickets per Order: total customer service contacts divided by orders fulfilled. This is broader than the return request rate: it captures every friction point, including delivery queries, sizing confusion, damaged goods, and fulfilment failures. A rising tickets-per-order trend predicts NPS deterioration and return rate increases before they appear in the numbers.

Product Review Score Trend: declining average review scores across Trustpilot, Google, or your own product pages predict return rate increases and NPS drops with a four to eight week lag. A move from 4.6 to 4.1 over two quarters is a leading indicator, not background noise.

Post-Purchase CSAT: post-purchase satisfaction scores measure the immediate transaction experience. A sustained drop signals fulfilment, product quality, or packaging issues before they show up in return rate.

First-to-Second Order Rate: the proportion of first-time buyers who make a second purchase within 60 to 90 days. This is the single most predictive signal of long-term LTV. A declining first-to-second order rate means customers are trying the product and not coming back. No CRM programme fully compensates for that. It moves six to ten weeks ahead of repeat purchase rate and cohort LTV.

Time Between Orders: the average number of days between a customer's first and second purchase. Tracked at cohort level, a lengthening time between orders is an early signal of weakening product habit before it shows in repeat purchase rate. It also sets the baseline for replenishment email timing. If your average is 42 days, deploying win-back flows at 30 days means you're treating customers as lapsed before they are.

Why most operators don't have this view

None of this is conceptually complicated. The problem isn't understanding what the metrics mean. It's that the metrics live in different places.

Outcomes sit in your finance system. Drivers are split between your marketing platform, your 3PL, and your CRM. Signals are scattered across your ad accounts, email platform, and support inbox.

Nobody connected them by default. That's not a personal failing. It's just how the tools were built. Each system was designed to track its own slice of the business, and nobody built the layer that joins them.

The operators who build this visibility usually do it one of two ways: they hire an analyst to pull it together manually in a spreadsheet, which lags and breaks whenever something changes, or they find a tool that connects it for them.

The second option is what Agenie is built for. Ask it which products are actually profitable after returns. Ask it what's happening to your CAC this week. Ask it whether your Tier 3 signals are pointing anywhere concerning. It connects across your stack so you don't have to.

Which metrics matter most by stage

Not every metric deserves equal attention at every stage.

$1M to $10M: can you make money on each order?

At this stage, survival is the primary question. The business needs to prove it can acquire customers profitably enough, retain enough of them, and fulfil each order without eroding the margin it just earned.

Prioritise: Contribution Margin per Order, CAC Payback Period, Gross Margin, Return Rate, First-to-Second Order Rate.

If these are healthy, you have a business. If they're not, more revenue makes the problem bigger, not smaller.

$10M to $50M: are the unit economics scaling?

Now the question is whether the model holds as you grow. Efficiency should be improving. If it isn't, you're scaling a problem.

Add: LTGP:CAC, Inventory Turnover, Email/SMS Revenue Contribution, On-Time Delivery Rate, EBITDA.

$50M+: is the business compounding?

At this scale, the cost of not connecting the three tiers is measurable in millions. The businesses that survive this stage are the ones that anchored every major decision in Tier 1 Outcomes while optimising in Tier 2 Drivers.

Add: Free Cash Flow, cohort LTV analysis, Days Inventory Outstanding, Sell-Through Rate by category.

A word on cohort LTV

Aggregate LTV is almost always misleading.

It averages across cohorts acquired at different times, in different channels, at different costs. The number can look stable even when recent cohorts are materially weaker than older ones.

What you need is cohort LTV: the lifetime value of customers acquired in a specific month or quarter, measured at consistent time points (three months, six months, 12 months).

If your Q4 2023 cohorts are worth less at six months than your Q4 2022 cohorts were, that's a clear deterioration in the quality of customers you're acquiring. No aggregate metric will show you that. And it usually points to something specific: a new channel mix, a product launch, a change in promotional strategy.

That's the analysis that actually changes decisions. Aggregate LTV just tells you what already happened.

The connection that changes everything

The operators who get this right aren't tracking more metrics. They're tracking the right ones together and watching how they connect.

When Tier 3 Signals start moving, they adjust before the Tier 2 Drivers shift. When Tier 2 moves, they diagnose it before it hits Tier 1 Outcomes.

That connection is what most tools don't make easy. The data sits in separate systems with different definitions and different update cycles. The relationships between metrics stay invisible unless someone is actively pulling them together, usually in a spreadsheet, usually too late.

If you're building that visibility and want to see how Agenie connects it across your stack, reach out. Happy to show you what it looks like in practice.

Georges

Key Takeaways

• Three tiers: Outcomes tell you if the business is healthy, Drivers tell you why, Signals tell you what's coming next.

• Read the framework top-down to diagnose a problem. Read it bottom-up to predict one before it arrives.

• Most operators live in Tier 2. The biggest gains come from connecting all three.

• Contribution Margin is not the same as Gross Margin. CM2, which subtracts variable marketing spend, is the number that tells you whether growth is actually profitable.

• Return Rate (Tier 2) is the damage already done. Refund and Returns Request Rate (Tier 3) is the warning, typically three to six weeks earlier.

• Aggregate LTV hides cohort deterioration. Track LTV at cohort level or the signal is invisible.

• The reason most operators don't have this view isn't sophistication. It's that the data lives in separate systems that don't talk to each other.

Frequently Asked Questions

Q: What are the three tiers in the DTC metrics framework?

Tier 1 (Outcomes) covers business health: Net Revenue, Gross Margin, Contribution Margin, EBITDA, Free Cash Flow, and NPS. Tier 2 (Drivers) covers the levers that explain why Tier 1 moves: acquisition, retention, commerce, operations, and inventory metrics. Tier 3 (Signals) covers leading indicators that move before Drivers shift and Outcomes deteriorate. Most DTC brands spend nearly all their time in Tier 2.

Q: What is contribution margin and why does it matter?

Gross margin tells you what's left after paying for the product. Contribution margin tells you what's left after you've also paid for everything that touched the order: landed COGS (product plus inbound freight and duties), fulfilment, 3PL fees, packaging, payment processing, returns handling, and any affiliate or influencer commissions on the transaction. That's CM1. CM2 goes one step further and subtracts total marketing spend at a blended level: paid media, agency, affiliate commissions beyond CM1, and any marketing cost that scales with growth. CM2 is the number that tells you whether growth is actually profitable. The most common errors: using catalog COGS instead of landed COGS, double-counting discounts that are already netted in revenue, and trying to compute CM2 per campaign when the metric only holds up blended.

Q: What's the difference between Return Rate and Refund and Returns Request Rate?

Return Rate (Tier 2) is the percentage of orders that have been returned and processed. It's a lagging measure of damage. Refund and Returns Request Rate (Tier 3) is the rate at which customers are requesting returns before those returns are processed. It's a leading indicator, typically moving three to six weeks before Return Rate and net revenue are affected. Tracking both separately is how you see the problem forming rather than just measuring it after the fact.

Q: Which DTC metrics matter most at early stage versus scale?

At 1M to 10M, the priority is proving the unit economics work: Contribution Margin per Order, CAC Payback Period, Gross Margin, Return Rate, and First-to-Second Order Rate. From 10M to 50M, the question shifts to whether the model holds at scale: add LTV:CAC Ratio, Inventory Turnover, Email/SMS Revenue Contribution, and EBITDA. At 50M+, Free Cash Flow and cohort LTV analysis become critical because aggregate metrics can hide deterioration that's already compounding.

Q: Why is aggregate LTV misleading?

It averages across cohorts acquired at different times, in different channels, at different costs. Even as recent cohorts weaken, the aggregate can look stable. Cohort LTV tracks customers acquired in a specific period and measures their value at consistent time points: three, six, and 12 months. That's where deterioration shows up early.

Q: How do I actually get visibility across all three tiers?

The honest answer is that most operators can't, without significant manual work, because the data sits across Shopify, Meta, Klaviyo, their 3PL, and their finance system. Connecting it manually through spreadsheets works but lags and breaks. Agenie connects across your stack so you can ask questions across all three tiers in real time, without the overhead.