Most successful companies sell products and services in a model in which they try to create long-term durable relationships with their customers. From simple examples like deodorant to complex products like network infrastructure, companies want to continue to sell you more products and services. Since the cost of acquiring a new customer is much higher than working with an existing customer the ability to successfully drive more revenue out of your existing customers is critical.
Correctly anticipating customers need is a key to success. Responding to customers needs is great, but what if you could predict the products and services they were most likely to buy and preemptively suggest them. This is exactly the model we see now on e-commerce sites like Amazon , which has seen predictive analytics and recorder technology drive 30% of their business. That’s billions of dollars of revenue generated, in large part, by presenting customers with the opportunity to purchase products that they already wanted ( whether they knew it or not)
The types of recurring business generally fall into three camps:
- Novel Products
Consumables are the easiest to conceptualize and implement. Printer companies have been incenting people to buy new ink cartridges through coupons and other direct messaging for years. Many channel partners drive a reasonable chunk of their revenue from anticipating when their customers will need new ink/paper etc. for years. In fact, the problem is so straight-forward and repeatable that many MSP’s have resources dedicated directly to making sure that any of the “low hanging fruit” of consumables gets serviced.
It gets more difficult when we talk about upgrades and responding to growth in our customer base. How do we know when IT departments will be moving towards replacing routers or legal firms will need to buy more furniture. Sure, it’s great when they reach out to us, but if we could anticipate their needs, it would undoubtedly be better and more profitable.
How do we do it? Can we apply some of the technology that underpins Amazon and the like’s recommenders and such? The answer turns out to be yes, kind of. There are a number of hurdles that need to be overcome including limited data, the role of the partner in any indirect sales process, and the difficulty of dealing with the fact the many customers will purchase a few items from us instead of the hundreds they might purchase from Amazon over a year. In addition , we must consider the nature of B to B purchases vs. B to C purchases.
That said, it does turn out to be possible to reliably predict the likelihood of a partners customer to purchase a given product within a specific window of time. We’ve figured out a way to use various other data elements ( which are publicly available for most businesses) to determine how similar customers are to each other and the product sets they have. This lets us “fill in the blanks” and gives us a much more robust data set.
One of the key wins here which isn’t so obvious is the ability to predict how likely it is that a current customer will purchase a “novel” product. That is, not an upgrade, not a refresh, but something entirely different than they have purchased in the past. Think of the famous example where more diapers are sold when they are placed next to the beer. Novel indeed!
What does this mean for vendors with channels? We can now seed our channel partners with leads into their existing customer bases and allow them to build revenue at a significantly lower cost than they would if the needed to “hunt” a new customer every time. What partner doesn’t want new leads?