Not All Visitors Make Great Customers – Moving Beyond Conversion Rates, Part 2

Jared Spool

May 21st, 2012

Moving Beyond Conversion Rates:

Part 1: Avoid Ratios for Metrics
Part 2: Not All Visitors Make Great Customers (this)
Part 3: Visitors Are Not All The Same
Part 4: Campaigns Are Where Conversion Rates Shine
Part 5: Measuring Money Left On The Table


A few years back, the executives at the electronics retailer Best Buy made a hard decision. They decided to become less attractive to a large group of their customers.

These customers were frequent purchasers, so you’d think they’d be someone Best Buy would want to court. However, these customers were driving up Best Buy’s costs and making it hard to serve others.

These customers would do crazy things, like buy a product on sale, use it for a few days, then return it, only to buy it back when the store put it out on a half-price open box sale. They’d get the credit for their original purchase at full price, but then buy the product again at half price.

Finally, the executives at Best Buy had determined these customers were hurting their ability to deliver quality service to the customers who wouldn’t play these games. That’s when they made the decision to stop serving a piece of their customer base.

In the first part of our series on moving beyond conversion rates, we looked at why ratios like conversion rates make bad metrics. Today, we look at a base assumption behind measuring the conversion rate of a web site: that every site visitor is desirable as a customer.

Not All Visitors Make Great Customers

Why wouldn’t we want to sell to everyone? Everyone drinks Coca Cola. Everyone loves McDonald’s. Why wouldn’t everyone want our product or service?

The problem is different customers come with different costs. Successful businesses focus on the most profitable customers and, at some point, draw a line on the customers they don’t want to serve.

Costs for customers happen in different ways. An unhappy customer may call support and want the support representative to help them make things right. That’s great if it’s something the product is supposed to do, but if it’s something the product was never intended for (and can’t do well), then that customer will never be happy and the support call is just added expense.

Unhappy customers tell others when they’re not happy. (Search my Twitter stream for mentions of United Airlines and you’ll see how this works.) It’s not hard to find reviews of mobile apps or Amazon products where the complaints are about things the product was never intended to do. These negative reviews scare away other customers.

For any product or service, there’s a group that shouldn’t be a customer or patron. It’s as important for the marketing and sales efforts to disqualify these customers as it is to entice the ones who will love the product or service, buy more without hesitation, and get all their friends to try it.

This is another place where optimizing a design for a conversion rate gets us into big trouble. Conversion rate says that every visitor should be converted into a purchaser.

However, if the marketing and sales components of our designs are doing well, we should intentionally see some visitors leave without a purchase. If we use a raw metric, that would make our conversion rate go down. Normally, we see a decreasing conversion rate as a bad thing, but in this case it’s really good. How do we account for that by charting the metric over time?

If we could filter the visitor denominator in our conversion rate metric to only focus on those who are the ideal, high-margin customers we’re seeking, then we’d have a lot more useful data. But, without a way to filter them out, the metric becomes filled with noise and substantially reduces its value.

In the next installment, we’ll look at how the visitors to our site aren’t as homogeneous as we’d like to believe, thereby reducing the helpfulness of a single conversion rate to designers.


Moving Beyond Conversion Rates:

Part 1: Avoid Ratios for Metrics
Part 2: Not All Visitors Make Great Customers (this)
Part 3: Visitors Are Not All The Same
Part 4: Campaigns Are Where Conversion Rates Shine
Part 5: Measuring Money Left On The Table

4 Responses to “Not All Visitors Make Great Customers – Moving Beyond Conversion Rates, Part 2”

  1. Roger Dooley Says:

    How true. In my paper catalog days, we tracked returns by customer, and sometimes dropped customers who ordered often but returned an unusual percentage of orders. We also occasionally found customers who were frequently problematic and consumed large amounts of time for our phone and supervisory staff. They were dropped from our mailings, which usually solved the problem without actively irritating the customer. That’s a bit harder with websites that the customer can always find.

    When one of those customers did attempt to place an order, we solved the pushback problem by telling the customer, “Despite our best efforts, we find ourselves unable to meet your high standards of performance. Rather than continue to disappoint you, we suggest you try another vendor, *insert hated competitor*.” Almost always worked!

    (My Twitter stream is littered with United comments as well. They have chose the “stony silence” approach to social engagement.)

  2. Kathy Sierra Says:

    Dear almost everyone on the Internet,
    Please read this (and the previous) post.

  3. Sveta Says:

    I wonder about things you buy online such as clothes and shoes that you cannot try on unless you order them? Sometimes it’s not a customers fault because various retailers have different size charts. It’s easier to try on those things on brick and mortar stores, but those items are sometimes available online only. What can be done to ensure that things fit customers without trying them on first?

  4. Alexis Brion Says:

    Totally agree. This is happening in many companies, specially in the ones the marketing department is too focused on the number of leads and not in the actual quality of them.

Add a Comment