One of the best things about artificial intelligence is that it processes oodles and oodles of historical AND real-time data.
SO MUCH DATA.
On the fly and at lightning speed.
Couple that with its unique predictive abilities, and you get actionable insights quicker and easier than ever.
Some of my favorites?
Likelihood to Purchase (LTP)
Likelihood to Repeat (LTR)
Likelihood to Churn (LTC or Churn Rate)
As described below, these three metrics are worth adding to your arsenal.
LIKELIHOOD TO PURCHASE (LTP)
Artificial Intelligence/Machine Learning has taken Likelihood to Purchase (LTP) to new levels by predicting who is most likely to buy with pinpoint accuracy. Full stop.
You can calculate Likelihood to Purchase for your online and offline efforts. (“Offline” = things like catalogs and direct marketing pieces.) Some marketers set up LTP as a hierarchical ladder or buckets (you often see this with sales teams), and others give every individual a precise score. Do whatever you’ll use and/or get value from. There’s no need for a fancy scoring system to the 9th decimal point if you only have a few set ways to react. (Ex: Group #1 gets an email. Group #2 receives an outbound call. Group #3 gets an email and a catalog.)
Some companies use LTP specifically to segregate all the folks who will NOT buy from them. This, of course, saves them time and money. It can also help with placement/deliverability. (Ad and SPAM algos have their own scores too. It’s beneficial to be on the same page.) Other companies use purchasing metrics to predict who is most likely to return their purchases or who will be a “problem” customer.
Likelihood to Purchase (LTP) has always been one of my very favorite metrics. It’s such a good indicator of how your marketing programs are working (or not.) It’s essential for folks in eCommerce and lead generation because it allows you to figure out which of your customers are low-hanging fruit and which are edible apples on the ground. Both juicy and VERY lucrative. Plus, it leaves clues about where the rat is getting stuck in the snake.* If you’re not tracking this yet, start. Yesterday.
Read more about Likelihood to Purchase here.
LIKELIHOOD TO REPEAT (LTR)
One-to-Two-X (now known as Likelihood to Repeat) looks at which of your customers will buy again and approximately when. It’s beneficial for companies who run a lot of Shopping and Social campaigns, where the customers are often fickler.
LTR has been around for years. That’s why traditional marketers still refer to it as One-to-Two-X or 1-2x. B2B and B2C companies look at their incoming new buyers and score them for future potential. Will they buy again? If so, how long will it take? This allows you to laser-target your marketing efforts (and monies) to every individual’s precise buying timeline. The more sophisticated users of LTR also look at how often individuals will buy over their lifetime.
Artificial Intelligence has helped metrics like LTR to skyrocket. After it assigns a score to a record indicating their likelihood to purchase again, it can follow a specific personalized process for the lead. For example, if I just bought from you and The Machine assigns me a score that I won’t buy again for two years, you could knock down the number of emails/SMS and other communications you send me from 5-7 a week to 1 every 10 days and reduce the number of mailings.
On the other hand, if The Machine says I will buy again within 60 days, you may want to ramp up the frequency and channel(s) of contacts. The Machine will follow through with whatever foundational system you’ve established and continuously improve it based on what it sees is and isn’t working. After you’ve built a solid foundation? You can let The Machine invent new and improved strategies for converting 1x customers to multi-buyers.
Marketers with sophisticated LTR programs often allow The Machine to individualize offers. So, Customer X just bought from you today, and you want to know what happens next. The Machine will figure out how long it will take before Customer X buys again and what offer(s) will make them buy faster or spend more money. Yes, that’s right. Each person gets their very own personalized conversion program!
LIKELIHOOD TO CHURN (AKA CHURN RATE OR LTC)
No matter how comprehensive your marketing efforts are, some folks will go into an Inactive status. They may stop buying from you; they may stop opening your emails or do something else to show their lack of interest. The good news is that you can prevent a big chunk of the churn from ever happening.
How? Knowing your Churn Rate. Unlike in years past, when marketers figured out their LTC rate in batches a few times a year, AI has made Likelihood an ongoing process. This metric is like looking into a crystal ball to see your future with each customer. The Machine determines who will churn when, and then it figures out how to prevent them from leaving you. By individual. What creative/messaging does Customer X need? What offer(s) do they need? How many contacts will it take? And so on… The Machine can figure out the plan, and it can implement the plan. (Or at least initiate the process.) Yowza! It’s crazy, I know.
Is this a fortune teller’s crystal ball, or is this a prediction you can take to the bank? That depends on how you set up and train your Marketing AI system. You’ll be golden if you set up a solid foundation/base (with training and testing protocols) and monitor it carefully, with human oversight.
This works well for buyers and inquiries. It is also successfully used in identifying folks who will likely not renew a Loyalty Club Membership or continue their Automatic Delivery Program. Plus, it’s good for figuring out who will cancel a subscription or unsubscribe from your newsletter. You determine where you want to use it and how you want the results implemented.
As with most Artificial Intelligence and Machine Learning projects, your success will rely on your data, foundation, and ability to disrupt things when they need a little reset. It’s easy (and uber common) to want to do ALL the things at once, but first, just track the basics of each of the Likelihoods above. Then, after you know what you’re seeing (and agree that the numbers are correct), start to put some simple systems in place. Keep working with The Machine, so it knows your best practices and processes. As it analyzes more and more data – including your real-time results – it will become better and stronger. Keep the training wheels on till you’re confident it can go on its own. It’s a process, it takes time, but it will be well worth it in the end.
Have questions or comments about the three Likelihood metrics? Tweet @amyafrica.com or write firstname.lastname@example.org
*People have squawked at me for using this analogy for over a decade. If you have something better, I’m happy to try it out. Although I’ve got to say, nothing gets a crowd to wake up quite like saying “snake” in the middle of a presentation. Freud would have a field day with Marketers.
A Down-and-Dirty Definition for Marketers. (Read more about these here.)