Did you really acquire that customer? At what cost?
A fresh look at customer acquisition in online retailing. Industry professionals talk about product management funnels involving intent, research, consideration, buying, repeat buying, recommending etc. in retail and other online businesses. In my book, you have either acquired a loyal customer or you haven't acquired any. While you ponder, do wonder about what is the right metric that captures the cost of customer acquisition. Hint: It's definitely not same is total marketing spend divided by number of users who placed orders.
Customer acquisition and its cost are some of the most talked about things in marketing and rightly so. They are also things that are optimized metrics. But is it really what we think it is? If we look at the most common interpretation, customer acquisition is assumed to start and stop at getting one transaction done by a user. If you run an online store and convert a window shopper to a first time shopper (on your store) with the help of one or more of the following - ads, spam (most advertising is spamming anyway), growth hacking, viral marketing, deals, discounts etc. it is widely accepted that you have acquired a customer! Is that really so? Is life that simple?
IMHO, it ain’t that simple, especially if the first transaction of your customer is unprofitable! And that usually is the case. The first transaction of the customer you have supposedly acquired tends to be lossy more often than not, after taking the acquisition cost into account. You still incur that loss in the hope that there is a lifetime value to the customer and she will turn profitable for you after many transactions. But is the lifetime value really there? It turns out that most customers either don’t have a huge lifetime value or the retailers fail to capture it. It’s a function of customer sat which itself tends to be a combination of customer service, delivery time, price, quality, look and feel, durability, brand etc.
As a customer I need to feel happy about all these to count as a satisfied customer. Is one transaction enough to judge all these satisfaction attributes positively? One transaction may be enough to complain. But it’s not enough to feel delighted, unless the delivery executive surprises me pleasantly by walking my dog in addition to delivering the goods I bought. In other words, customer delight takes time (customer dissatisfaction may or may not) and more than one transaction. And what is the meaning of customer acquisition if you haven’t acquired a delighted or at least visibly satisfied customer. It’s only such customers who tend to have a lifetime value. Also it does take some getting used to before visiting your store proactively (with a need or without one) becomes a habit. So how many transactions does it take? Can you engineer enough transactions to create delight? What is the real cost of acquisition?
Let’s address the questions one at a time.
How many transactions does it take to ensure delight? Let’s get the basics right. If there are issues with your customer support policy, communication, delivery time, quality control, the genuineness of brands etc. more transactions will only ensure more dissatisfaction. Assuming you’ve got these sorted, your store may need 3 or 5 or 7 transactions before your average customer feels something of a delight or at least satisfaction - enough to put your store in the consideration set and proactively visit your store when she thinks of buying something in the categories you deal with. This varies from customer to customer, and store to store. That’s the reason the operative word is “average customer”. How many transactions is that for the “average customer”? I don’t have a number. But I have a methodology.
Anecdotally, the more a customer has experienced shopping in your store, the more likely you are to be in her consideration set when she is in need of something that you sell.
Because of the variability between stores (differences in categories, brand recall, ease of shopping etc.), there is hardly any alternative to surveying an unbiased sample of customers. The caveat is however, customers may not know the number either. You have to use oodles of creativity to design such surveys. For example, you could ask the customer when (after how many transactions) she first typed your store name on the browser or launched your app without prompting by ads, emails, texts, push notifications etc. If your sample size is decent and you have avoided things like self-selection bias, surveys don’t lie.
If you have a large data science team, without more serious problems to solve, you can make these inferences from user behavior rather than be asking questions. Maybe data science can tell you that a regular window shopper (regardless of the actual transaction frequency) is a loyal customer after all! Since our data science, people are busy solving simpler (but more useful) problems, we at recosenselabs.com did a good old survey (sample size less than 1000). The headline graphic illustrates the results. If it's not self-explanatory, do leave a comment!
Naturally, the cost of customer acquisition is the cost of getting these many transactions done – in terms of advertising, deals, cash back etc.
Is there a way to engineer that many transactions. Of course there are multiple ways. It has gotten slightly harder after Google decided to crash your promotional email campaigns. And Android does allow users to shut push notifications for each app!
So how will you do this? Here’s a simple check list.
1) Piggyback promotional messages on transaction (not just financial transaction) emails.
2) Offer enough cashback to do the next transaction; Nothing is as magnetic as funny money sitting in the customer’s account with the store! This is a variable cost; If it’s hurting your margins, that’s a good problem to have; It means the cashback is working
3) Personalize push notifications and other communication.
4) Merchandise your storefront as if it’s a bespoke suit – tailor made for each customer.
5) Avoid spamming with generic/frequent emails and push notifications; Almost nothing is as harmful to the dignity of a customer as being spammed en masse!
6) Have an element of surprise in your communications and offers.
7) Use retargeting intelligently – Guess what the customer might be most likely to buy given her shopping and widow shopping habits so far; Don’t insult her intelligence by showing me the same products that she shopped / window shopped; While retargeting, do remind the user about how much funny money she has lying in her account with your store!
8) Monetize abandoned carts by using price drop alerts and such communication.
9) Keep in mind – I, as your customer, am lazy, stingy and stupid; That’s almost my fundamental right. Don’t try to inspire me to work hard for a discount or solve Fermat’s Last Theorem.
10) Slowly reduce the cashback and other financial incentives (only those intended to get the required number of transactions) after the customer has done enough transactions to keep you in the consideration set for one or more categories.
11) Branding ads on television, youtube etc. can increase (forcible) the presence of your store in the customer’s life and hence increase the brand recall of your store when there is a need; Do these if you have the budgets; If your category is too niche, don’t bother with this.
Can you do all this and still be sustainable? Frequently, marketers talk in terms of cost of customer acquisition vis-à-vis the average lifetime value of the customer.
All those calculations will fall apart if you replace cost of customer acquisition by cost of loyal customer acquisition. But that’s the right metric. If you haven’t acquired a loyal customer, it is almost as good as not having acquired any customer at all. See if your business model makes sense even after having made this replacement. In a maturing investment market, that’s’ what would convince investors (if they are willing to listen to begin with) your business is a sustainable one. Remember, despite your best efforts, if the competition has more staying power, none of these metrics are worth a penny. That’s for you and/or your investors to assess before getting in.