Definitive Guide to Running an e-commerce Business in 2021
Importance of thinking about customers rather than orders (and setting a Target CAC)
How much to spend on marketing?
by Al Taylor
Updated June 28th 2021
We’ll start super simple – but build up quickly:
It’s important to understand your contribution margins – how much of your revenue you get to keep and where the money you spend is going
When someone places an order then you have to send the customer the goods ordered.
There is a cost to you of servicing that order:
- Sales tax;
- COGS (Cost of Goods Sold – the amount you paid for the goods purchased);
- Cost to store, pick and ship the goods;
- Customer service;
- Card processing fee; and a
- Share of Marketing costs.
Let’s look at these costs a little more closely:
The Unit Economics of an Order
Let’s take an example. A customer spends £120 on your site. The customer paid £114 for products and £6 for shipping the goods.
In the UK there is 20% Sales Tax (VAT – value added tax) on both the products and shipping. So once the sales tax has been removed the order value is £100 (£95 products and £5 shipping)
As discussed, not all the £100 of revenue goes into your back pocket.
There were costs to service that order.
Below is an example breakdown of the costs associated with the order.
I’ve grouped packaging, pick and pack, card processing, customer service into ‘cost to process’.
CM3 (Contribution Margin 3)
CM3 is the most important metric – it is what is left over from an order to pay for the overheads of the business.
The light green at the bottom of the bar chart is your CM3. In this example when the customer spends £100 (excl tax) you get to keep £15 (15% of the order value) for your back pocket – 85% are costs associated with winning and servicing the order.
This is what you want to optimise as it’s this that allows you to pay salaries and make a profit.
CM1 (Contribution Margin 1)
CM1 (Contribution Margin 1) is what’s left after the cost of the goods has been accounted for.
CM1 = Product and Shipping revenue minus product cost of goods sold.
This is an indicator of your core product margin.
You can make CM1 higher by increasing prices; increasing shipping fees; reducing discounts; purchasing the goods for less; or changing the mix of products purchased into higher margin products.
CM2 (Customer Margin 2)
CM2 (Contribution Margin 2) is what’s left after all costs associated with fulfilling the order have been accounted for (think of it as all costs except marketing).
Improving CM2 is about more efficient warehouse, shipping, packaging, customer service and card processing costs.
The aim is to optimise your CM3 – the money left to pay salaries, office costs and other overheads. Whatever is left after your overheads is your profit.
Your total CM3 in a month is the average CM3 per order multiplied by the number of orders.
A very high CM3 per order but with very few orders = low total CM3.
A very low CM3 per order and lots of orders = low total CM3.
You want a strategy that delivers a good CM3 per order AND lots of orders = high total CM3.
Price vs Conversion Rate
Let’s use an example to show how lower prices can make you more CM2
Bob run’s an e-commerce store that sells T-shirts.
Each T-shirt costs Bob £5 to manufacture (Cost of goods of £5) and he sells them for £30 each (excluding tax).
Over the course of the month Bob has 10,000 visitors to his website and the conversion rate to order is 1% (delivering 100 orders).
The average order is for 2 T-shirts (@£30 each) and there’s a £5 shipping fee. So the average order value is £65.
The cost to ship is £5 and the cost to process each order (pick, pack, store, packaging, card processing etc) is £5.
The unit economics (or unit stack) looks like as below. Bob makes £45 CM2 per order (what’s left before marketing costs).
His total CM2 from his 100 orders in the month = £45 * 100 = £4,500
Kate runs an identical website but she prices her T-shirts at £20 each.
Because of the lower retail price the conversion rate is much higher at 3%, so she has 300 orders a month (10,000 x 0.03 conversion).
Kate’s unit stack looks like this:
So the average order value is £45. After COGS, shipping and costs to process the order she makes £25 CM2 (versus Bob’s £45 CM2 per order when the T-shirts cost £30 each).
However the total CM2 (CM3 before marketing costs) across all orders in the month is £25 * 300 = £7,500 (versus £4,500 for Bob whose T-shirts are priced at £30)
So in this example the higher conversion rate means that the lower priced T-shirts actually generate more overall CM2 for Kate than Bob (66% more in fact – great!).
Marketing Spend and Conversion Rate
Bob’s lower conversion rate means his marketing spend per order is much higher than Kate’s
Let’s now consider the marketing cost.
For both stores 5,000 monthly visits come from ‘free sources’ such as Organic Search and 5,000 visits come from paid sources (such as Facebook ads or Google Ads) at an average cost of £0.50 a click.
Therefore the total marketing spend in the month for both Bob and Kate = Clicks x Cost per Click = 5,000 * £0.5 = £2,500
Bob’s store with marketing spend (£30 T-shirts)
For Bob, the blended marketing spend per order (this means the marketing spend spread across paid and free orders) is £2,500 / 100 = £25 per order
So Bob’s CM3 per order is £20 and his total CM3 across all orders = £20 x 100 = £2,000
Kate’s store with marketing spend (£20 T-shirts)
For Kate, the blended marketing spend per order is £2,500 / 300 = £8.33 (because Kate’s conversion rate is 3 times higher than Bob’s, her marketing cost per order is three times lower than Bob’s).
So Kate makes £16.67 CM3 per order.
Across all 300 orders Kate makes £16.67 * 300 = £5,000 total CM3.
So Bob makes £2,000 CM3 (with his higher priced T-shirts) while Kate makes £5,000 CM3.
Kate makes 66% more CM2 than Bob but 150% more CM3 (and CM3 is the key metric as it’s what goes into your back pocket).
Blended vs Paid Cost per Order
Applying Bob’s Marketing spend against only orders from paid sessions shows he is losing money on each paid order. Kate’s higher conversion rate means she is making money.
So far we have looked at blended marketing cost (applying the marketing cost across all orders, free and paid).
Remember however that half the orders came from ‘free’ visits and half came from paid visits.
Let’s attribute the marketing cost to just the paid visits and orders
For Bob’s store the total marketing spend is still £2,500 (5,000 clicks @ £0.5 per click).
The 5,000 paid clicks have a 1% conversion rate which delivers 50 orders.
The marketing spend per paid order is £50 (£2,500 spend / 50 paid orders) which means each paid order has a CM3 of -£5 (i.e. you lose £5 on every order).
That doesn’t sound good.
Let’s look at Kate’s Store where we sell T-shirts at £20 each.
The £2,500 marketing spend delivers 150 orders, so the marketing spend per paid order is £16.66.
This delivers a CM3 per order of £8.34 – so we make money on each paid order – much better.
Cost per Order vs Cost per Click
It’s dangerous to focus on Cost per Click as it has no sense of quality of click (likelihood the person clicking will convert)
Let’s look at Bob’s marketing spend again.
He spends £2,500 to generate 5,000 visits @ £0.5 a click.
With a conversion rate of 1% Bob gets 50 paid orders with a marketing spend of £50 per paid order – meaning he makes -£5 on each order (CM3 = -£5)
Let’s look at the marketing spend in more detail. Bob runs three different Google Ads campaigns:
The ‘Beach T-shirts’ campaign delivers the most clicks (3,500) and the lowest Cost per Click (£0.41) – however, because the conversion rate is low (0.57%) the marketing spend per paid order is £71.75 driving a CM3 per order of -£26.75 (he loses £26.75 on each order – yikes!)
The ‘Organic T-shirts’ campaign however has a cost per click twice as high (£0.84) yet because the conversion rate is over 4 times better at 2.5% the marketing spend per order is much less at £33.60 – giving him a much healthier £11.40 CM3 per order.
This is an example that focusing on cost per click is a bad idea – The campaign with the much more expensive cost per click is the one that drives profitable orders!
So by only running the Organic T-shirts Campaign (turning off the other two campaigns) Bob can move his CM3 per paid order from -£5 to +£11.40.
Across free and paid channels Bob now gets 75 total orders in the month but his total CM3 goes up from £2,000 to £2,535 as we have eliminated the unprofitable orders (£45 CM3 for each of the 50 ‘free’ orders and £11.40 for each of the 25 paid orders).
Every click costs a different amount and has a different likelihood of converting.
Focusing on Cost per Order is much better than Cost per Click
Cost per Order = Number of Clicks x Cost per Click / Conversions
Cost per Order is a much better metric to focus on because it encapsulates conversion and cost per click – i.e. it’s OK to pay a high cost per click if there is a high conversion rate.
Similarly low conversion rate can be OK if the cost per click is low.
Cost per click gives you no insight into the quality of the traffic coming to the site. Having a really low cost per click that drives people who don’t purchase is a waste of money.
Perhaps Bob can get the ‘Beach T-shirts’ campaign to work without changing the T-shirt price or on-site experience.
To get the it to work Bob needs to target the people that are more likely to convert and/or get the cost per click down for that audience.
It might be that people on a particular device, using particular keywords, particular Facebook audience are more likely to convert: high conversion is good.
Similarly having your ad appear only 60% of the time for a relevant search versus 90% of the time will reduce the cost per click: lower cost per click (at the same conversion rate) is good.
So going after a tighter audience and bidding less aggressively should deliver a much lower cost per order on the Beach T-shirts campaign.
For paid orders to be profitable the average cost (marketing spend) per order must be less than the average CM2 (when Marketing spend per order = CM2 then CM3 is zero).
So targeting a cost per order that is lower than your CM2 means you will on average have profitable paid orders.
Click Through Rate (CTR)
NOTE: Click Through Rate (CTR) is another metric that people focus on – but there are caveats with this too.
CTR is what % of people that see your ad (an impression) click on it.
Improving CTR means more volume which is great if your campaign delivers profitable orders but not so great if your campaign delivers unprofitable orders.
The incremental volume might be people less likely to convert too.
So before focusing on CTR make sure your campaign is profitable.
A side benefit is that a high CTR typically reduces your CPC as ad platforms reward ads that get a response from the audience (in Google Ads you get a higher quality score).
However, if people bounce (only look at one page) or have low dwell time (don’t stay long) when they click on your ad then that’s a negative signal for Google.
So remember, CTR is about volume not quality. For a campaign that has fantastic conversion and CM3 then you want as many people as possible to click on the ad.
However, if the ad has terrible CM3 then you don’t want people to click on it until you’ve improved it’s conversion rate.
Repeat orders are great – and if you have loyal customers you can afford to spend more on the first order knowing more orders are likely to come
Some of Bob and Kate’s customers will make repeat orders. Repeat orders are great!
Someone who has purchased from Bob’s store before is likely to have a higher conversion rate on subsequent visits (great!)
A repeat purchaser is also more likely to have come via a free channel, such as direct, organic search or email (great!)
The average customer to Bob’s store makes 2 orders in their lifetime.
Let’s take one customer. The first order is from a paid channel and the second order via email. Both orders are for two T-shirts with an order value of £65.
As we saw earlier, the marketing spend per paid order on Bob’s store was £50. With CM2 of £45, Bob loses £5 on the first order (CM3 = -£5).
On the second order there is no marketing cost, so the CM3 = CM2 = £45
So across the lifetime of the customer Bob makes -£5 + £45 = £40 CM3
So when looking on an order basis Bob was losing £5 per order but across the customer’s lifetime, Bob makes £40 CM3 on that customer.
How to square this? 50 of the 100 orders a month Bob gets are from ‘free sources’ – where did these orders come from? A lot of them would have come from customers placing repeat orders where Bob had acquired the customer through a paid channel.
Life-Time Value (LTV) & Cost of Acquisition (CAC)
If you understand how much money you are likely to make from a customer over their lifetime (say 2 years) then you can work out how much to spend on acquiring them
To make money (CM3) from a customer we need the total CM2 across all that customer’s orders to be more than the marketing spend on that customer (across all orders).
The sum of all a customer’s orders’ CM2 is called the Life Time Value (LTV) or sometimes the Customer Lifetime Value (CLTV).
Another way to think about this is that Lifetime Value equals the total revenue from a customer (called the lifetime Revenue or LTR) minus all the costs of getting them the orders (except marketing).
The lifetime value (LTV) of a paid customer at Bob’s store = CM2 of first order + CM2 of repeat orders = £45 + £45 = £90
So if Bob wants to make money (CM3) he needs the Cost of ACquisition (CAC) of a new customer (plus any repeat marketing spend) to be less than the LTV.
So Bob wants the total marketing spend on a customer to be less than £90 over the customer’s lifetime or he won’t make any money.
If Bob has an average CAC = LTV = £90 then he acquires lots of customers but makes no money on any of them (For simplicity I’m rolling any repeat marketing spend into CAC).
If any customer costs Bob more than £90 to acquire than he will lose money on that customer.
We can say the maximum CAC Bob should be willing to pay is £90 which equals the LTV (why would you ever want to lose money on a customer?)
At an average CAC of £90 Bob makes no CM3 as the total marketing spend = total CM2.
If Bob does no paid advertising then his CAC = 0 but he gets no paid customers so again he makes no money.
So Bob wants his CAC to be somewhere between zero and the LTV. A reasonable target CAC is half his life-time value (LTV/2).
What CAC should you aim for? Let’s understand the demand curve better:
Currently Bob runs three campaigns.
If Bob just ran the ‘Organic T-shirts’ campaign (shown in green) he would get 25 orders at a cost per order of £33.60. To get another 5 orders he runs the ‘colourful T-shirts’ campaign (blue) but they cost £44 each. To get more orders he runs the ‘Beach T-shirts’ campaign (red) which gets him another 20 orders but they cost £71.75 each.
For Bob’s business (and all businesses) the demand curve is not elastic.
This means Bob can’t keep acquiring more orders for the same cost per order.
Within a campaign the cost per order will typically go up if he chases more orders (e.g. he gets 25 orders from the organic T-shirt campaign at £33.60 a order. If he wants to get 30 orders from that campaign then it may cost him £40 an order).
This means that each additional customer you acquire is slightly more expensive than the previous customer.
When you have an average CAC of say £30 that doesn’t mean it cost you exactly £30 to acquire each customer.
Some customer would have been acquired for less than £5 and some for more than £60 (and this is all simplification anyhow, let’s not forget that the customer technically only cost you their click but you have to bundle in all the clicks/views from the people who didn’t purchase).
The high cost to acquire the next customer gets lost when averaged out across the cost to acquire all customers.
Say you acquire 100 new customers in a month at an average of £10 each.
You then push marketing harder and now drive 200 new customers at an average of £15 each.
The additional 100 customers did not cost £15 each to acquire but £20 each (on average). [100 customers @ £10 each and 100 customer @ £20 each ⇒ 200 Customers @ £15 each].
That is again a simplification – The 101st customer may have cost £13 and the 200th customer £30.
In theory you want to keep acquiring customers until your last customer’s CAC = LTV (under the premise that it’s worth having a customer that only makes you say £1 over the customer’s lifetime than not having that customer).
There are two problems with this, firstly, pushing your marketing spend such that your last acquired customer’s CAC = LTV pushes up the cost of acquiring the cheaper customers too (the logic being that by pushing your campaigns harder, making them appear more frequently pushes up the cost per click across the whole campaign) – i.e. pushing harder is not just about acquiring incremental customers, it will push up acquisition cost on your cheaper customers too.
Secondly, it’s very difficult to measure what the CAC of your last customer is!
Bob’s Paid orders include new and repeats
Another complication is that Bob’s Paid Orders will include some new customers but also some repeat orders.
Repeat customers tend to convert better than new customers (Bob might have had 500 paid visits from existing customers driving 10 repeat paid orders and 4,500 paid visits from non-customers leading to 40 new customers. Because of the higher conversion rate on the repeat customers, Bob’s marketing spend per repeat paid order was 500 x £0.5 / 15 = £16.66 but the marketing spend per new order (CAC) was 4,500 x £0.5 / 40 = £56.25)
How Much to Spend (LTV/CAC)
LTV/CAC helps us work out the right amount to spend on each customer
The simplicity of LTV and CAC is it tells whether you should be spending more or less
So Target CAC tells you how much you should spend on acquiring a new customer.
If your actual CAC is more than your Target CAC then you should reduce your marketing spend. If your actual CAC is less than your Target CAC then you should increase your marketing spend.
Not all Customer Groups have the same LTV
You need to decide how to group your customers and track LTV. You may find that customers in different geographies have very different LTVs (perhaps because of different AOVs, repeat frequencies).
Or there might be different LTVs by category (customers buying sports shoes might be worth more than those buying sports socks) and different LTVs by channel.
Track LTV on a too detailed level and it becomes overwhelming (plus if you are too granular there is not enough data to see trends).
The aim is to make LTV higher
By making LTV higher, each customer becomes more valuable and you can therefore spend more to acquire each customer (your Target CAC goes up).
The higher your Target CAC the more new customers you win and the business grows fast as you get more new and repeat customers.
The best way to have higher LTV is to improve the number of repeat orders. This is one of the reasons that Amazon is so focused on customer experience – ease of discovery, clarity on delivery etc – as it helps bring people back to the store to order again.
Promotions and discounts can help, as can focusing on acquiring email addresses so you can email your customers and leads to tempt them back to your store to purchase.
Making the Average Order Value higher helps as well. This can be done through ‘add-on’ purchases – nudges to buy something additional and surprisingly through good merchandising and promotions.
Product and company reviews give customers confidence to purchase.
Quite often when you give the customer a discount or better value they actually spend more (something we didn’t factor in to the Kate’s store example above – in reality she may have sold on average 2.5 T-shirts per order vs 2 for Bob).
The aim is to make your marketing spend be as efficient as possible at your Target CAC
This simply means that you want to get as many new customers as you can at your Target CAC.
The key to achieving this is to get your Google Ads and Facebook Ads to trade off each campaign you run against each other – letting the platforms work out where to spend that next pound of spend to get the best return. This means optimising to orders (or revenue) and using a portfolio to manager campaigns or ad sets.
How to set-up your Google Ads and Facebook Ads to get the most orders at your Target CAC
For you to accurately track your CAC you also need to be able to identify different types of spend. Here are some examples:
Lead Generation (get someone’s email address)
When calculating Paid CAC we want to focus on the Purchase spend. When calculating Blended CAC we should include all marketing spend.
Find out more here:
How to Name Your Google Ads and Facebook Ads Marketing Campaigns
How to Track LTV and CAC
Mapflo can automatically track your LTV and CAC by channel and website – here’s the logic we use:
How to Track your LTV and CAC and Calculate your Target CAC
You can find out more about tracking and creating a reporting dashboard in this article:
How to set up a dashboard to track the performance of your e-commerce store
Think about CAC as what you pour in and LTV as what you get back.
If you have an LTV / CAC of 4 then for every £100 you spend on acquiring a customer you get £400 of CM2 back over the customer’s lifetime ⇒ £300 CM3 (£400 CM2 less £100 marketing).
The higher the number the better return you get on your investment, but the fewer customers.
Blended LTV / CAC.
If you calculate your LTV / CAC across all your channels (free and paid) we call this Blended LTV/CAC (blended as you are combining your ‘Free LTV / CAC’ and your ‘Paid LTV / CAC’).
For your free channels the CAC may be zero (though you can attribute costs to lead generation, content generation, technical SEO, email creation etc).
Paid LTV / CAC
For Paid channels we can set a Target LTV / CAC. There is an optimal LTV / CAC that drives the most absolute CM3 (where CM3 per customer x number of customers is highest).
We think LTV / CAC = 2 is probably not far off that optimum – if your distribution of orders versus CAC is a normal distribution (see diagram) then your Target CAC is half your Maximum CAC (Maximum CAC = LTV) – with a perfect distribution the last customer you acquire you would break even on.