Sales slow down. Production has room. The instinct is to tell marketing to generate more demand. Sometimes that’s exactly right. Sometimes it just sends more traffic into the constraint that was already limiting the business.
This article gives owners of direct-to-consumer product companies a way to tell the difference — six questions, asked in order, that show whether marketing is the thing holding growth back or whether something else in the business is.
Marketing is probably your bottleneck when the business converts visitors well, makes acceptable money on each order after acquisition, can fulfill more orders and can fund them — and simply isn’t getting enough qualified buyers.
If any of those other conditions is already failing, more traffic is unlikely to be the first thing to fix. It tends to make the real constraint more expensive. The job is to find the binding constraint — the one actually stopping profitable growth right now — and solve that first.
In this article
What is a growth bottleneck, exactly?
A growth bottleneck is the current binding constraint on profitable growth: the one limit that, if you relieved it, would let the business grow — and that nothing else would get past.
The idea comes from operations. The Theory of Constraints, developed by Eliyahu Goldratt, starts from the observation that a system’s output is set by its constraint, and that the first step in improving it is to identify that constraint before acting. Improve anything else and the system barely moves. Relieve the constraint and the whole system speeds up — until a new constraint appears somewhere else.
That last part matters for marketing. A bottleneck isn’t permanent. The constraint in March may not be the constraint in September, which is why the question “is marketing our bottleneck?” needs asking more than once.
Something that looks worse than you’d like. Low conversion, a high acquisition cost, soft order volume.
The limit actually stopping profitable growth right now. Relieve it and the business grows; leave it and nothing else helps much.
Most businesses have several weak metrics at once. Only one of them is usually binding. Low conversion doesn’t automatically make conversion the growth constraint. A high acquisition cost doesn’t automatically mean advertising is the problem. The work is deciding which weakness matters most right now.
Why don’t low sales automatically mean low demand?
Because flat sales is a symptom, and several different constraints produce the same symptom.
When orders stall, the most visible lever is marketing, and the most visible marketing lever is spend. But order volume is the output of a chain: people have to find you, choose to buy, receive something the business can make and ship, and — in many models — come back. A break anywhere in that chain shows up as the same flat line.
- DemandNot enough qualified buyers are arriving.
- ConversionThey arrive, but too few buy.
- EconomicsOrders are coming, but growth is being bought at a loss.
- RetentionFirst orders were supposed to pay back later, and they aren’t.
- CapacityStockouts, backlogs or cash limits are capping what can be sold.
Only the first of those is solved by generating more demand. For the others, more traffic can make the problem larger before it makes it smaller.
How do you find the constraint that’s actually limiting growth?
Ask six questions in order. Each one covers a different part of the business, and each has a different answer to “what should we do next?”
- 01DemandAre enough qualified potential customers entering the system?
- 02ConversionAre we turning the qualified visitors we already get into orders?
- 03EconomicsDo those orders leave enough money after acquisition?
- 04RetentionIf the model depends on repeat purchase, is it actually happening?
- 05CapacityCan inventory, fulfillment, operations and cash absorb more orders?
- 06DecisionGiven all of that: scale test, hold or fix first?
The order is deliberate, but it isn’t a strict ranking. A severe problem further down — a stockout on your best seller, or a month that loses money — can answer the question before the earlier layers are even checked. Treat the framework as six lenses on one business, not a checklist to be completed top to bottom.
Are enough qualified buyers actually reaching us?
Demand is the constraint when the rest of the system is healthy and ready, and the missing input is simply enough of the right people.
Traffic and demand aren’t the same thing. Traffic is visits. Qualified demand is people with a realistic chance of buying this product at this price — the right need, the right budget, the right timing. A campaign can double traffic without adding much demand at all, and more low-quality visits mostly show up as a falling conversion rate.
What a genuine demand constraint tends to look like
- Capacity and inventory are available, and the operation could handle more orders well
- People who do arrive buy at a rate that’s normal for your business
- New orders leave acceptable contribution after acquisition
- New-customer volume, not order economics, is what’s short
- Campaigns are limited by budget or by how many relevant people they can reach, while their results stay acceptable
When those line up, marketing may well be the bottleneck — and the right response is usually a controlled increase in demand, not a leap. Which channel makes sense depends on whether the demand already exists and needs capturing or has to be created; Meta Ads vs. Google Ads for e-commerce covers that choice.
Are we turning the visitors we already have into orders?
If qualified people are already arriving and too few are buying, more traffic mostly buys more abandoned visits.
Conversion problems rarely announce themselves as conversion problems. They show up as “the ads stopped working” or “acquisition got expensive.” The causes usually sit on the site or in the offer:
- Product pages that don’t answer the questions a buyer actually has
- An offer, price or bundle that isn’t competitive for this audience
- Missing trust signals — reviews, guarantees, a real business behind the store
- Shipping costs or delivery times that surprise people at checkout
- Mobile experience or checkout friction
- A mismatch between what the ads promise and what the page delivers
That last one cuts both ways: sometimes the “conversion problem” is really a traffic-quality problem, and the fix is better targeting rather than a better page. Either way, the diagnosis comes before the budget.
On benchmarks: published average conversion rates vary enormously by category, price point, traffic source and how “conversion” is defined. They can offer rough context. They can’t tell you whether your conversion is the constraint. Your own history — the same pages, the same traffic sources, before and after — is far more useful.
Do the orders we’re creating actually leave enough money?
Revenue growth isn’t automatically profitable growth. Marketing can produce more orders while the economics under them get worse.
The number that matters here is contribution after acquisition: what each order leaves once product cost, the variable costs of fulfilling it — shipping absorbed, pick and pack, packaging, payment processing, discounts and returns — and the cost of acquiring the customer have come out. Contribution margin is the portion of revenue left to cover fixed costs and profit after variable costs; it isn’t the same as gross margin, and it isn’t net profit.
- Orders
- 2,000
- Contribution before acquisition $30 an order
- $60,000
- Acquisition cost
- −$22,000
- Contribution after acquisition
- $38,000
- Orders
- 2,400 +20%
- Contribution before acquisition $30 an order
- $72,000
- Acquisition cost
- −$41,000
- Contribution after acquisition
- $31,000 −18%
The 400 extra orders cost $19,000 to acquire — $47.50 each — against $30 of contribution each. Every extra order lost about $17.50. The dashboard shows growth; the business kept less.
When more customers make the economics worse, demand isn’t the first constraint. The question becomes pricing, product mix, order value, discounting, fulfillment costs or acquisition efficiency. A strong-looking return on ad spend doesn’t settle it either: return on ad spend measures attributed revenue against media, not what the business keeps — the subject of our article Why Do We Have a 5x ROAS but No Profit? For how contribution sets a sustainable budget, see how much an e-commerce business should spend on marketing.
Do we depend on repeat purchases that are actually happening?
Retention is the constraint when the business needs future orders to justify what it pays for the first one — and those future orders aren’t showing up, or nobody can see whether they are.
There are two legitimate models. In one, the first order is expected to pay for itself; repeat purchase is welcome but not required. In the other, the company deliberately accepts thin or negative first-order economics because customers reliably come back — consumables, replenishment, subscriptions. Both can be sound.
Common with durable products bought rarely. Low repeat purchase isn’t a flaw here — it’s the nature of the product. Acquisition has to work on the first order.
Rational only when repeat behavior is measured, by the month customers were acquired, and payback actually arrives on a timeline the business can fund.
The risk sits in the second model when payback is assumed rather than measured. If first orders lose money and the recovery depends on purchases nobody has confirmed, scaling acquisition isn’t automatically the answer — it scales the loss while the evidence is still missing. The fix may be measurement, the offer, or the post-purchase experience rather than more demand.
Can the business responsibly absorb more demand?
For a physical-product company, marketing success doesn’t stay in marketing. Every order lands on inventory, fulfillment, customer service and cash.
This is the seam most marketing advice skips, and it isn’t a new problem. Research published in MIT Sloan Management Review found that when the demand side and the supply side of a company are disconnected, businesses end up selling excess product below market rates or losing sales to inventory shortages — often running separate financial, marketing and operations plans that don’t agree. More recent work from MIT’s Center for Transportation & Logistics describes e-commerce growth adding pressure across demand forecasting, inventory and fulfillment at once.
Capacity also means more than machines. What it covers depends on the kind of product company you run:
Machine time, labor and shifts, raw materials, floor space, production scheduling, lead times and quality — plus fulfillment and cash.
Purchase-order timing, supplier lead times, stockouts and excess inventory, 3PL throughput, warehouse space and cash tied up in stock.
Stockouts are the clearest example of why this matters to marketing. In a large field study at a direct-mail retailer, researchers found that a stockout hurt more than the missing item: customers cancelled other items in the same order and bought less over the following 13 months. The authors caution that the exact size of those effects won’t transfer to every business. The direction is the point: demand you can’t fulfill can cost you customers you already had.
Shopify frames the same trade-off from the inventory side — forecasting is how a store balances overstock against stockouts without tying up unnecessary capital. Marketing is one of the biggest inputs to that forecast.
Signs capacity is already the constraint
- Stockouts on products you’re actively advertising
- Backlogs, longer ship times or rising overtime
- Customer-service volume climbing faster than orders
- Cash stretched between paying for inventory and paying for acquisition
None of this makes Coast333 an operations consultancy, and the answer isn’t always to hold demand back. Companies sometimes build demand ahead of capacity on purpose — preorders, waitlists, a planned expansion, a retail launch. The principle is coordination: demand generation should be planned alongside inventory, cash, capacity and fulfillment rather than discovered by them.
What if we can’t trust the numbers?
Then the right move is usually not to pick whichever dashboard looks best. It’s to hold off on any large scaling decision until the gap is understood well enough to decide responsibly.
Most owners live with several versions of the truth: Meta’s numbers, Google’s, Google Analytics, the store platform, the agency report and the accounting system. They rarely match, and a mismatch doesn’t necessarily mean something is broken. Google’s own documentation notes that differences between Google Analytics and Google Ads are common and can occur even with a correct setup — they count and date conversions differently.
The deeper issue is that attributed results aren’t the same as caused results. In a study of 15 large advertising experiments at Facebook, the observational measurement methods common in the industry often failed to match the results of randomized experiments on the same campaigns. And field experiments at eBay found that returns from paid search were a fraction of what non-experimental estimates suggested, largely because frequent buyers who would have purchased anyway received much of the credit.
None of that means you need perfect attribution to make any decision. You don’t. It means confidence should match evidence. The practical anchors are company-level: net sales, new-customer counts and contribution from your own books. When platform reports and those numbers tell a broadly coherent story, you can act with more confidence. When they don’t, keep changes small, or hold, until they do.
Lower measurement confidence doesn’t forbid a decision. It shrinks the size of the decision you can responsibly make.
So what should we actually do next?
Every path through the framework ends in one of three decision categories: scale test, hold, or fix first. They’re directions, not guarantees.
- Scale test
- The business appears ready to test more demand — in controlled increments, watching the same gates as volume grows.
- Hold
- Don’t materially increase demand yet. Conditions call for stability, observation or better information first.
- Fix first
- Another constraint should be addressed before buying materially more traffic.
StartIs qualified demand what’s in short supply?
Are conversion, economics and capacity all healthy?
If yes, demand may be the constraint. Consider a controlled demand test.
If any of them isn’t healthy, work down the gates below — they’ll show which one.
Demand may not be the problem. Work down the gates below to find what is limiting growth.
-
1
Is conversion healthy against our own history?
Yes — continueIf noFix firstConversion may be the first constraint to investigate.
-
2
Are customer economics healthy after acquisition?
Yes — continueIf noFix firstFix the economics before scaling acquisition.
-
3
If first orders lose money, is the repeat purchase that recovers it actually measured?
Yes, or first orders stand alone — continueIf noHoldMeasure payback before scaling.
-
4
Can inventory, fulfillment, cash and operations absorb more orders?
Yes — continueIf noHoldCapacity is currently constraining growth. Coordinate before adding demand.
-
5
Is measurement confidence adequate for a material decision?
Yes — continueIf noHoldImprove measurement before a material scale decision.
Every gate passes. Test more demand incrementally — and keep checking the same gates as volume grows, because the constraint will move.
What does this look like in practice?
Six patterns cover most of what owners run into. They’re illustrative — real businesses rarely fit one exactly — but each shows how a single binding constraint changes the answer.
| Case | Demand | Conversion | Economics | Retention | Capacity | Measurement | Likely reading |
|---|---|---|---|---|---|---|---|
| A | Short | Healthy | Healthy | Not deciding | Available | Adequate | Scale test Demand may genuinely be the bottleneck. |
| B | Healthy traffic | Falling | Acceptable per order | Not deciding | Available | Adequate | Fix first Don’t assume more traffic is the answer. Investigate conversion. |
| C | Orders rising | Not deciding | Poor after acquisition | Not deciding | Not deciding | Adequate | Fix first Marketing is producing volume; economics are the constraint. |
| D | Not deciding | Not deciding | First order loses money | Required, unmeasured | Not deciding | Weak on payback | Hold Understand retention and payback before scaling. |
| E | Strong | Not deciding | Not deciding | Not deciding | Stockouts, backlog | Not deciding | Hold Marketing isn’t the limiting factor right now. Capacity is. |
| F | Healthy | Healthy | Healthy | Healthy | Available | Platforms and books don’t reconcile | Hold Confidence is too weak for an aggressive scale decision. |
Case F is the one owners find most frustrating, because everything looks fine. But “everything looks fine” is only as reliable as the numbers saying so. A modest, well-measured test is often the right step there — not a large one.
What does it look like when marketing really is the bottleneck?
It happens, and it’s worth recognizing quickly — because waiting too long leaves capacity and inventory sitting idle.
The pattern of a genuine demand constraint
- The business can fulfill substantially more orders, well
- Contribution after acquisition is acceptable
- Conversion is healthy for your own history
- Inventory is available for what you’d promote
- Cash can fund more acquisition before it pays back
- Existing customers alone don’t fill the available capacity
- Qualified traffic and new-customer volume are the missing input
- Small, controlled increases in demand keep producing acceptable economics
Read these together. Any one alone proves little; the combination is what points to demand.
The last signal is the most important, because it’s the one that tests the others. If each step up in spend keeps delivering acceptable economics, demand is the constraint and the business is ready for it. If results start to slip, the constraint has probably moved — and the framework starts over. When demand is the constraint, the next question is where it will come from; a Competitive Marketing Analysis looks at how competitors are competing for customers online and where there may be room.
What should you fix before spending more on ads?
Whatever is already binding. The point isn’t to spend less — it’s to solve constraints in the right order.
| If you’re seeing | More traffic tends to | Look at first |
|---|---|---|
| Conversion deteriorating | Buy more visits that don’t buy | Pages, offer, checkout, traffic quality |
| Weak contribution after acquisition | Scale weak unit economics | Price, mix, order value, costs, acquisition efficiency |
| Stockouts or fulfillment strain | Create backlog and cancelled orders | Inventory position, purchase timing, throughput |
| Customer service overloaded | Lower the experience for every customer | Staffing, self-service answers, root causes of contacts |
| Weak repeat purchase the model relies on | Grow first-order losses | Payback measurement, post-purchase experience |
| Measurement nobody trusts | Enlarge a decision no one can evaluate | Reconciling reports to company results |
| Cash or working capital pressure | Spend cash before it comes back | Payback timing, inventory commitments |
Some of these are marketing’s to fix — traffic quality, landing pages, acquisition efficiency. Others sit with operations, finance or the product. Good marketing people will point them out rather than work around them. For how to judge whether your ad accounts are being actively managed at all, see how to tell whether your e-commerce ad account is actually being managed.
What does solving the wrong bottleneck cost?
Usually more than the spend itself — because effort on the wrong constraint leaves the real one untouched.
- If conversion is the constraintMore traffic buys more abandoned sessions.
- If economics are the constraintMore orders scale weak contribution.
- If capacity is the constraintMore demand creates backlog, cancellations and a worse customer experience.
- If measurement is the constraintMore spend increases the size of a decision nobody can evaluate.
- If demand really is the constraintHolding back leaves productive capacity and inventory unused.
That last line is why this isn’t an argument for caution. Under-investing when demand is the constraint is as real a mistake as over-investing when it isn’t. The discipline is the same either way: find the binding constraint, solve it, and then look again.
What should I ask the people running marketing?
You don’t need to become a media buyer. You need a short set of questions that connect marketing to the business — and answers in business terms.
- Demand
- Are we actually short of qualified buyers?
- Conversion
- Are the people already arriving buying at a reasonable rate for our own history?
- Economics
- What does a new order leave after variable costs and acquisition?
- Retention
- Are we relying on repeat purchases we’ve actually measured?
- Capacity
- How many additional orders could we fulfill well next month?
- Cash
- Can we fund the acquisition and inventory before the money comes back?
- Measurement
- Do company results and marketing reports tell a coherent enough story?
The single most useful question
Which constraint would still stop us if marketing produced 20% more orders tomorrow?
Illustration: a company shipping 2,600 orders a month with practical capacity for about 3,000 would need 3,120 to absorb a 20% jump — 120 more than it can handle well. Before that extra demand arrives, capacity becomes the binding constraint. If nothing would stop you, demand probably is the constraint.
The answers also tell you something about the people giving them. A team that can explain results in terms of contribution, capacity and cash is managing marketing as part of the business. A team that can only answer in platform metrics may still be good at the platforms — but you’ll be making the business decisions without them. If you’re weighing a change, when to change e-commerce marketing agencies covers which problems a new partner can and can’t fix.
This framework comes from operating inside the problem. Coast333’s founder spent roughly eight years inside a direct-to-consumer manufacturer handling about 10,000 orders a month — across operations and, eventually, as Head of Marketing — where every marketing decision sat alongside production, inventory, fulfillment, margin and capacity.
So is marketing your growth bottleneck?
Only when insufficient qualified demand is the constraint actually preventing the business from growing profitably.
Marketing’s job isn’t to maximize demand. It’s to help create the right amount and quality of profitable demand for the economic and operating system the company is trying to run — coordinated with inventory, cash, capacity, fulfillment and the company’s strategic goals. Sometimes that means spending more, sooner. Sometimes it means holding until another part of the business catches up. The discipline is knowing which, and being willing to hear either answer.
The same framework also gives you a way to judge marketing advice. A recommendation that starts with a channel, a tactic or a budget number before anyone has asked what’s limiting the business has skipped a step. Diagnosis comes first.
The system can only grow as fast as its current constraint. Find it before you fund it.
Frequently asked questions
How do I know if marketing is the bottleneck in my business?
Check the other conditions first. If visitors convert well, orders leave acceptable contribution after acquisition, you can fulfill more orders and fund them, and your numbers are trustworthy enough, then a shortage of qualified buyers is probably what’s limiting growth — and marketing is the bottleneck.
If sales are down, should I increase ad spend?
Not automatically. Find out why sales are down first. If demand fell while everything else stayed healthy, more spend may help. If conversion, economics, inventory or fulfillment changed, more spend usually makes that problem more expensive.
How do I know if I need more traffic or a better conversion rate?
Compare conversion with your own history for the same pages and traffic sources. If it’s stable and healthy while qualified traffic is short, traffic is the gap. If qualified traffic is arriving and conversion has fallen, fix conversion before buying more visits.
Can marketing be working even if the business isn’t growing profitably?
Yes. Marketing can deliver the orders it was asked for while product cost, discounts, fulfillment or acquisition cost leave too little behind. That’s an economics constraint, and more volume can make it worse.
What should I fix before increasing my marketing budget?
Whatever is currently binding: conversion problems, weak contribution after acquisition, unmeasured payback the model depends on, stockouts or fulfillment strain, cash pressure, or measurement nobody trusts. If none of those applies, you may be ready for a controlled increase.
How do I know if capacity is limiting growth?
Look for stockouts on promoted products, backlogs, lengthening ship times, rising overtime, customer-service volume growing faster than orders, or cash stretched between inventory and acquisition. Ask how many more orders you could fulfill well next month.
Can poor retention make customer acquisition look too expensive?
It can make it genuinely too expensive. If your model relies on repeat purchases to recover what a new customer costs and those purchases aren’t happening, acquisition costs that looked affordable aren’t. If your product rarely gets repurchased by design, judge acquisition on the first order instead.
What if Meta, Google and Shopify all show different results?
That’s common, and each may be accurate by its own rules. Anchor on company-level numbers — net sales, new customers and contribution — and treat platform reports as supporting evidence. The less they reconcile, the smaller the scaling decision you should make until you understand why.
Should I stop marketing if operations are constrained?
Usually not entirely. Holding demand steady, shifting spend toward products with inventory, or using waitlists and preorders can all make sense. The goal is to coordinate demand with what the operation can fulfill, not to switch marketing off.
How often should I reassess the growth bottleneck?
Whenever something material changes — a season, a product launch, a price change, new capacity, a big budget change — and at least monthly. Relieving one constraint usually moves the bottleneck somewhere else, so last quarter’s answer may not hold.
Sources & further reading
- The performance of the theory of constraints methodologyMabin and Balderstone, International Journal of Operations & Production Management, 2003.
- Integrating Supply and DemandMIT Sloan Management Review, 2015.
- 2026 State of Supply Chain Omnichannel Report (summary)MIT Center for Transportation & Logistics.
- Measuring and Mitigating the Costs of StockoutsAnderson, Fitzsimons and Simester, Management Science, 2006.
- Contribution Margin: What It Is, How to Calculate It, and Why You Need ItHarvard Business Review, 2017.
- Conversion discrepancies between Google Analytics and Google AdsGoogle Ads Help.
- A Comparison of Approaches to Advertising Measurement: Evidence from Big Field Experiments at FacebookGordon, Zettelmeyer, Bhargava and Chapsky, Marketing Science, 2019.
- Consumer Heterogeneity and Paid Search Effectiveness: A Large-Scale Field ExperimentBlake, Nosko and Tadelis, Econometrica, 2015.
Scenarios and dollar figures in this article are illustrative, created to show the reasoning. They aren’t industry benchmarks or client data.
