Last time we started a conversation about growth and left it off with a brief discussion about digging deeper into customer behaviors.
In particular, we were starting to look at the behavioral signals that indicate your best clients or simply the signals that show they’re about to become a customer. Either way, it’s valuable information to have.
To continue from where we finished last time, let’s turn one of my final comments from the last piece into a question: “What behaviors identify a person about to buy?” and “What behaviors identify potential ‘ideal’ clients who are already a customer – but haven’t transformed into that ideal customer quite yet?”
Order history helps
When we’re looking for behaviors that indicate what a great (“fanatical”, to use Rackspace’s lingo) customer is, one place to look is order history.
Your order history is rich in information that can help your detective work on buying signals and customer behavior.
I know, you want some examples of what you might look at. From gut feel, identify your five best customers, using whatever means you use to determine “best”.
Your business may not fit all of these questions I’m about to ask, so look at them in a way understand that they might require a slight adjustment. It’ll depend on what you sell, how you sell it and how rich your product line is.
- How often do they buy from you? In other words: What’s the average number of days between purchases?
- Does transaction size increase over time or does it shrink over time?
- Of the customers who buy everything you offer, do their purchase intervals or transaction sizes “look different” than everyone else’s (on average)?
- If your best customers don’t buy everything, what do they buy that no one else buys? Of those people, study the behavior prior to that particular purchase. What did they ask? What did they buy just before that?
Now…looking at the patterns that these “best” customers have established, what *existing* customers fit the early part of those patterns? These are the customers who are likely to join the “best customer group”. The difference is that you know the candidates in advance.
Not all of them will move into your best customer group, but in watching this process/movement, you’ll eventually learn what behaviors indicate that move.
Look at your interaction data for each class of customer. When I say “class”, I mean your best customers, your newest customers and so on. You need to look at each because you’ll need to be able to detect a behavior that occurs when a customer moves from one to another.
When they do that, they’re sending a signal. Your responsibility is to act on it.
InteractionsÂ include sales and support inquiries, price list requests, orders, email (including subscribes) and the like. Remember “guinea pigs” vs. “guinea pig” from last time? Your most important indicator of “I’m going to be a great customer” could be that subtle.
Interactions and order history indicate future behavior. Your best customers’ behavior is there and shows patterns along the road to “BestCustomerVille”.
By now you’re probably wondering about the not-too-standard things that I’m suggesting you observe. What about standard metrics? Where do they fit in?
Standard metrics, like the number of customers you have, the number of leads you have, the number you add each day/week/month, sales this month vs. last month and so on are certainly worth looking at, but remember that they primarily indicate *where you are*, not what you need to do to (re)produce those gains.
What do you learn from knowing that you have 1344 customers today and that you had 912 customers this time last year? Unless you look in the context of what resulted in the net gain of 432 customers, you learn little that allows you to reproduce that gain.
That’s what you want to know and repeat.
A “You are here” marker doesn’t help much if you don’t have a map. The behaviors we discussed are part of the map that shows you the path from new customer to great customer.
THAT is why we’re talking about behaviors. They are the invisible signals you have to detect.