Here's how a $179,865.82 increase in sales is actually a -$1,063.81 loss in profits when you crunch the right numbers.

As a company, the DTC society is almost brainwashing us into offering free shipping on all orders. "Conversion rate" is the talk of the town, and we have all read how shipping is the number one reason for abandoned carts.

But let me tell you once and for all: you don't have to offer free shipping. You shouldn't. It's losing you money.

By the end of this article, you will:

1. Understand why free shipping on all orders is losing you money.

2. Have access to a financial model where you can calculate the exact profit or loss from implementing free shipping on all orders.

3. Understand what to offer instead and how to make that "what" as profitable as possible.

4. Have a 20% off discount code for ShipScout, a Shopify tool you can use to A/B test all your "what" from step 3.

Let's dive in!

The Problems of Free Shipping

If you offer free shipping on all orders or have ever considered doing it, you probably argued that the free shipping would increase your conversions rate and hence your sales. And if you're not making "enough" money on the first order, it's a good play because of all the returning customers.

But it's not that simple. Here's why:

1. With free shipping on all orders, your customers lose the incentive to put more items in their baskets. Therefore, your average order value (AOV) decreases. If the increase in conversion rate is not enough to offset the decrease in AOV, you are losing revenue.

2. We're here to increase profits, not sales. What happens with profits when the number of orders increases (increased conversion rate) while the average order value goes down? Profits go down. Therefore, you have to include costs in your decision. Most people don't.

3. You most likely forget all the derived cost effects of an increase in conversion rate that are difficult to quantify.

Let's dive into each. If you want to run your own numbers as you read along, copy my Google Sheet here. I recorded a Loom here that explains how you use the sheet. I just realized that I forgot to talk about the output numbers in the Loom, but it should be self-explanatory. The "Revenue Delta" and "Gross Profit Delta" numbers are how much your revenue and gross profit increased/decreased after implementing free shipping.

Problem 1: Your average order value (AOV) goes down

Let's assume that you're currently offering free shipping above a certain threshold, but you want to test offering free shipping on all orders. According to ShipScout, the average store going from free shipping above ~$100 (our current free-shipping threshold) to free shipping on all orders increases the conversion rate (CR) by 13.97%. 

If I blindly consider CR alone and use numbers from one of my stores, a 13.97% increase in CR results in $179,865.82 more in sales. Not bad. That's a lot of money.

But what happens when I also consider the effect on my store's average order value? According to ShipScout, the average store with our free-shipping threshold sees a 6.18% decrease in AOV after implementing free shipping on all orders. After taking this into account, the increase in sales declines to $89,181.70. Still not bad, though!

Problem 2: Profits > Revenue

While the extra $89,181.70 sounds excellent, there is a problem. An increase in orders doesn't just bring more revenue, and it also brings more costs; shipping, packing, salaries, etc.

"But aren't we paying for all that in both scenarios?" Yes, you are right. But the problem is that you are paying all the costs of a 13.97% increase in numbers, but you are only getting margins to cover costs from the 6.9% increase in revenue. And even worse, as your average order value has decreased, you are making fewer margins on each order. In other words, your costs are a higher percentage of each order.

I've done the math to show you the effect of this. Without including costs in my calculations, my gross profits increase by $39,239.95 from implementing free shipping on all orders. But when I include shipping costs (incl. returns and packaging) and salaries as best as possible, the number drops to $21,160.20.

So the 13.97% increase in orders results in a 6.9% increase in revenue and only a 4.8% increase in gross profits.

Problem 3: You forget the derived cost effects

Alright. Our experiment still seems to increase our profits. But there's more.

What happens when you remove the friction of a shipping cost? Your customers lose the incentive to make sure that they want the product they're buying. I'm in the clothing business. If I remove shipping costs, my customers are more likely to buy a size they already know they don't fit. Or try a color they don't even like because it's at a discount.

So what happens? Returns go up. The problem with returns is that you pay the same costs per order as "normal" orders, but you don't get any revenue out of it.

You could say that by removing the friction of shipping costs, your customers have to pre-qualify less. They don't have to "want" to be your customer as much. So what happens? Your overall lifetime value of each customer goes down. I have plenty of customers who always return whatever they buy even though I make them pay for the shipping. If I remove the shipping, they will buy even more they end up returning.

But it's not just your direct, variable costs that go up. If you suddenly have to handle 13.97% more orders, you must also buy and hold more inventory. You need a larger warehouse. Your number of customer support tickets goes up. More orders bring both more costs that are difficult to quantify but also more risk.

Alright, get to the point, Mathias. How does this affect the numbers? If I increase returns by 10% and factor in and include an "unquantifiable cost" of 5% of total variable costs, which I honestly think is low, the gross profit declines to -$1.061,81.

Here's a visual representation of what's going on. The numbers are slightly different from above because I added VAT to all costs to compare apples with apples.

The Kicker(s)

Now, -$1.061,81 might not sound like much. And you might feel that I had to invent costs to get the gross profit to a negative.

But here's the kicker: We're a high AOV store. In theory, we should be able to afford to decrease our margins on each order to get more orders. For example, if we were a $60-AOV store, we would bleed money both with (-12,856.12) and without (-4,936.14) the "unquantifiable costs."

And let's say that we would be able to make even $10,000 in gross profits. Is that even worth the extra work and risk considering we're doing $1,6M?

And not only that. The CR increase and AOV decrease I used are averages from ShipScout. In reality, the CR increase will be smaller, and the AOV decrease larger. Why? Because it turns out there's an optimal shipping threshold that maximizes the trade-off between CR and AOV. So if you ran the above numbers comparing free shipping and your optimal free shipping threshold, you would most likely see free shipping bleeding money compare to the shipping threshold.

A note on returning customers

I mentioned in the introduction that returning customers will compensate for the possible loss on a customer's first order. I went down a rabbit hole last year and ran the numbers. Returning customers will not compensate for the loss on the first order. Read my rabbit-hole tweet here.

The easy way to think about it is if the profit per visitor is maximized with a free shipping threshold (which it is for most stores), the second visit will also be more profitable with a free shipping threshold.

Solution: Offer free shipping above your optimal threshold

Now that we've established that free shipping on all orders is not the way to go, we have to figure out what to do instead.

The first thing to understand is that we want to offer free shipping - above a certain order value threshold. The theory is that the threshold will incentivize your customers to put more items in their baskets. This incentive increases your AOV and consequently your bottom line profits.

The challenge is to figure out what your optimal threshold is. And notice that I write "your optimal" threshold. If you don't find your optimal threshold, you're leaving money on the table, and free shipping on all orders might look like it's more profitable.

Luckily, there's a tool for that: ShipScout. ShipScout lets you A/B/x test different free shipping thresholds. It works by dividing your traffic into different pools that get free shipping at the various thresholds you set. When you have enough data, you can see which threshold is the most profitable.

Note: I have no financial interest in ShipScout. I don't get any kickbacks, affiliate money, or anything. I just love their product so much that I reached out to one of the Co-founders, Ben, and asked whether they wanted to give something to my audience.

During my talk with Ben, he showed me case studies where the profit per visitor increased by $0.06 by setting the right threshold. That's $72,000 more in profits per year for one of my stores - just by setting a better free-shipping threshold. Here's how:

Step by step guide to finding your optimal free shipping threshold

Start by install ShipScout on your Shopify store using this link. By using this link, you're getting 20% off the ShipScout subscription. That's between $120-$240/year, so remember to use the link.

Go to the ShipScout app and start a new "Free Shipping Thresholds" test with the following setup:

• Variant A: $0.00. By setting it to $0.00, you include a free shipping variant in your test. This is your control variant as want to be sure that free shipping on all orders is suboptimal for your store.

• Variant B: If you already have a free shipping threshold, use this for variant B. If not, set variant by to around your average order value multiplied by 1.2. While this rule of thumb has some drawbacks, it's a fine start.

Now, you wait and collect data. When you have enough data, the interesting data is the "Profit Per Visitor" (PPV) you can find in ShipScout. We don't care about revenue, conversion rates, or average order value. We care about profits.

To get correct numbers, input your variable cost/per order (shipping, salaries, packaging) as well as your profit margin. With our current data, variant C is the best in my test. However, it's important that you have enough data. To establish when you have enough data to conclude anything, look at the "Statistical Significance" sections in the "Revenue Per Visitor" section. If you have > 95% confidence in the same variant (C), you should be good to go.

After you've established whether free shipping or your free shipping threshold (variant B) is best, it's time to optimize. Assuming that variant B gives you more profit per visitor, you simply start a new test with the variant B threshold as your new control variant, and then a higher and a lower threshold as variant B and C. If you find a new winner, repeat this exercise until you've established your optimal threshold.

An example of interpreting an experiment

To give you a better understanding of the dynamics, consider the following experiment I'm currently running on one of my stores. It's the same experiment as the above Profit Per visitor chart.

The experiment contains three variants, each with different free shipping thresholds:

• A: Free shipping above 700 SEK

• B: Free shipping above 800 SEK

• C: Free shipping above 900 SEK

Conversion Rate

Before looking at the numbers, I would guess that variant A converts much just from the fact that it has a lower threshold. However, while it's not significant, variant b is the winner so far.

Revenue Per Visitor

Which variant has the highest revenue per is harder to predict as it's a function of both conversion rate and the AOV. If I had to guess, I would guess that the variant with the highest CR also results in the most revenue per visitor.

However, that's not the case. It turns out that variant C actually results in more revenue per visitor even though it's converting at a seemingly lower rate. This tells me that the average order value of variant c is much higher than that of variant b. So the effect of getting customers to buy more to get free shipping dominates the decrease in conversion rate. The result is fewer orders, but higher profits per order.

Profit Per Visitor

Now, the most important metric. It's not surprising given the above result, but it's often easier to understand when presented this way.

In order to get the profit per visitor, you have to input an average shipping and fulfillment cost per order and our profit margins. After inputting my guesstimates, we see that variant c results in 0.51 SEK (~$0.06) more profit per visitor than the second-best variant. That's a lot of money with hundreds of thousands of visitors each year - just by increasing the shipping threshold.

The results are still not significant. But when they are, assuming that variant c wins, I will start a new experiment with variant C as the base case and then a new variant with an even higher shipping threshold - and repeat until the new variant doesn't perform better than the base case.