7 COD Mistakes Costing Shopify Merchants Real Money

The invisible problem of COD

Anyone can activate cash on delivery on Shopify in an afternoon. Releasit, EasySell or Shopify's manual method cover it. What almost nobody sees is what happens afterwards: how profitability silently degrades month by month until the business operates at a loss without realizing.

The seven mistakes that follow are what I saw (and committed) running a Shopify with COD in Spain. They're ordered by economic damage — the first can wreck all the others.

Mistake 1: measuring COD as a general average

Imagine an ad campaign with two ads. One has CPA of $50, the other $5. No one in their right mind would leave both active. You turn off the $50, scale the $5, done.

In COD, exactly the same happens. Some SKUs in your catalog confirm at 95% and deliver at 90%. Others confirm at 60% and deliver at 50%. When you only look at the general average, you don't know which is which. You end the month with 70% global delivery and "you don't know what to optimize".

The consequence: problem SKUs stay active, eating ad budget and reverse logistics, while good ones bear the cost. It's like leaving the $50 CPA ad on because the campaign average still looks fine.

The fix: measure confirmation and delivery per individual SKU, not aggregated. Filter products below average on either indicator. Remove them from the COD catalog or offer them with online payment only. The difference between operating at a mediocre 70% average or a depurated 88% defines annual profitability.

Mistake 2: not confirming phone at registration

COD is the only payment method where phone matters more than name or address. The name can have a wrong letter, customer is still identifiable. Address gets corrected by the courier on the spot. But a phone wrong by one digit means you can never confirm the order — that lead is lost before leaving the warehouse.

Typical COD audience in Spain is older than card-paying audience (40-65 years in many verticals). Typos are common. Add to that intentional fake phones from people who "fill forms out of curiosity", and you can lose 5-12% of leads just from this problem.

The fix: SMS OTP at registration. Customer receives a 4-digit code, types it into the form, only then the order completes. SMS costs $0.02 to $0.05 per verification. Cash on delivery without OTP is like accepting cards without validating CVV.

The form gets slightly longer. You accept that. The drop in fake phones from 8% to 1-2% pays for itself the first week.

Mistake 3: confirming late (more than 24 hours after registration)

The COD lead is impulsive. Customer decided in thirty seconds to complete the form because they didn't have to pull out their card. If the call center calls two days later, customer already cooled off the decision. Operational reports:

• Confirmation within first 4 hours: acceptance rate 85-92%
• Confirmation between 4 and 24 hours: 78-85%
• Confirmation between 24 and 48 hours: 65-75%
• Confirmation past 48 hours: drops to 50% or less

The fix: confirm same day as registration. If they registered at night, first hour of the next day. There's no excuse for "we'll call tomorrow" when talking about COD. Every extra hour between lead and confirmation costs conversion points.

And especially: the weekend is decisive. If your confirmation team only works Monday to Friday, you massively lose leads from Saturday night and all of Sunday. A part-time team confirming via WhatsApp on weekends pays for itself improving conversion 5-8 points.

Mistake 4: launching COD campaigns on Thursdays or Fridays

This is the most subtle and probably least known. The reason is in the logistics calendar, not in customer behavior.

Example in Spain with standard courier:

Thursday afternoon and Friday leads wait 4 to 5 days between confirmation and delivery. In COD, that interval is where the customer is lost: changes mind, sees the same product elsewhere, simply forgets the order. Rejection rate at delivery rises significantly on orders with more than 3 days of wait.

The fix: if you're launching a new campaign, scaling budget or testing a product, start on Saturdays or Monday mornings. Those days' leads reach the customer in 2-3 days max, within the window where decision stays hot.

Thursday afternoons and Fridays are valid for maintaining already-optimized campaigns — not for scaling or testing. It's an operational decision that doesn't affect ad targeting but drastically changes delivery rate.

Mistake 5: allowing COD on heavy or expensive-shipping products

COD scales poorly with heavy, voluminous or expensive-to-return products. Three reasons:

1. Courier charges re-attempts by weight — a 30 kg re-attempt costs double a 2 kg one
2. Customers reject large products when seeing them at the door — the friction of "didn't expect something this big, I don't want it" is much higher without prior economic commitment
3. Returning a heavy product to warehouse eats several dollars of reverse logistics per unit

In one operation I had with 40 kg dumbbells, delivery rate dropped to 50%. Dumbbell re-shipment cost $25 each time. The product was removed from the COD catalog after three months because the math didn't work.

The fix: filter COD catalog by weight and package dimensions. Products over 5 kg or with large dimensions (backpacks, voluminous electronics, furniture, fitness equipment) go to online payment only. Small, light products with medium AOV ($30-80) are the ones that perform best in COD.

Mistake 6: depending on a single COD app

If your only COD app goes down on a Saturday night with high traffic, you lose hours of revenue with no support team responding. Shopify App Store COD apps have a history of reported outages in 1-2 star reviews — including Releasit, the most popular.

The solution isn't moving to a "more stable" app because none guarantees 100% uptime. The solution is having two apps installed simultaneously, one active and one dormant ready to take over. Releasit as primary + EasySell free as backup costs $9.99/month and eliminates single point of failure risk.

Full comparison between the two apps and how to configure the dual setup is in another article: Releasit vs EasySell: Why Running Both Isn't Paranoia.

Mistake 7: calculating CPA dividing by leads instead of sales

The apparent CPA shown by Meta, TikTok or Google divides ad spend by leads (completed forms). In COD, that's deceptive because only part of those leads convert to real sales.

Concrete example from my operation:

• Ad spend: $1,000
• Leads generated: 100
• Apparent CPA from dashboard: $10
• Leads that confirm: 89
• Leads that receive on first attempt: 68
• Total delivered with re-attempts: 72
• Real CPA per delivered sale: $1,000 / 72 = $13.89

39% higher than apparent. And we still haven't added re-shipment costs, warehouse returns and call center operations. When you sum everything, complete CPA can be 50% higher than apparent.

If you scale budget thinking CPA is $10 when it's really $15, you can end the month paying more in ads than COD net billing. The complete explanation with cost table is in: COD in Shopify: The Real Funnel and Why Your CPA Is Miscalculated.

The fix: every time you look at the ads dashboard, mentally multiply apparent CPA by 1.3-1.5. That's the real cost. Scaling decisions get made on real CPA, not apparent.

The metrics you need to see every week

If your COD is serious, these are the five metrics you need to review weekly — not monthly, weekly — to avoid the mistakes above:

1. Confirmation rate by SKU (not aggregated)
2. Delivery rate by SKU (not aggregated)
3. Real CPA by acquisition channel
4. Post-COD net margin by SKU (after deducting reverse logistics)
5. Delivery rate by day of week of registration

Building this dashboard manually from Shopify + courier + ads requires 4-6 hours weekly of operations. Most operators don't do it because it doesn't scale.

This is what I'm building in CODprofit, a Shopify app specifically for COD operations analysis that automates these five indicators. Subscribe to the newsletter below for early adopter notification with launch discount.

Verdict

Frequently asked questions

Which of the seven mistakes is the most expensive?

Number 1 (measuring aggregated instead of per SKU). It hides the other six. If you have one operational change to make this month, it's building the confirmation + delivery + net margin report per SKU, not aggregated. With that visible, the other mistakes become obvious.

Is it worth outsourcing the confirmation call center?

Yes once you pass 200 orders/month. Below that volume you do it with one internal part-time person. Outsourcing earlier dilutes the sales script (generic call centers don't understand the product), and improving confirmation rate 10 points pays for the call center cost several times over.

How much is it worth investing to improve delivery rate?

Let's do the math. If your net margin per sale is $20, and you have 200 confirmed orders/month with 75% delivery = 150 sales, that's $3,000 of margin. Raising delivery from 75% to 85% is 20 extra sales = $400 extra margin monthly = $4,800 yearly. That justifies spending up to $3,000 on operational improvements (better courier, scripts, follow-up team) and still comes out profitable.

Do the seven mistakes apply equally in LATAM and MENA?

Mostly yes. Specific differences:

• In MENA late confirmation is punished even harder (shorter decision window)
• In LATAM courier weight on re-attempts varies enormously by country
• In India OTP is practically mandatory due to shared phone numbers

But underlying logic (measure per SKU, not aggregated; real CPA vs apparent; dual app setup) applies in all COD markets.

Why are Thursday afternoons worse than Tuesday afternoons?

Because of the week cutoff. A Tuesday afternoon lead confirms Wednesday, ships Thursday, arrives Friday. Three continuous days without weekend in between. A Thursday afternoon lead confirms Friday, ships Friday/Monday, arrives Tuesday. Four to five days with Saturday and Sunday in between where nothing moves and the customer can cool off. The weekend gap is what defines.