Happy Hour!

I mentioned in December that I was thinking of hosting a Happy Hour where you could ask me any questions you might have as well as having general discussions with the attendees. So let's try this ... I don't care if four people sign up or forty.

Happy Hour

• Thursday, February 13.
• 8:00pm EST / 5:00pm PST.
• 55 Minutes.

If you want to attend, send me an email (kevinh@minethatdata.com) and I'll forward you a link for Happy Hour.

Bring your favorite beverage - whatever that is, water is fine, something more interesting is fine as well.

I'll work on a topic to get us started.

Who's in?

+/-

In the NBA, players are occasionally evaluated on a "+/-" metric. If you come into the game with the score 42-34 and you leave the game with the score 54-44, your +/- is +2 (you were ahead by eight when you came in and you left leading by ten).

In Action Streams, +/- represents a change in customer state.

• A "+" is when the customer does something positive, like visiting your website or interacting with your community or watching one of your videos on YouTube or clicking through an email campaign.

A positive action requires a new Action Stream. If the customer has not purchased in eighteen (18) months and suddenly appears on your website, you suspend your normal email cadence with a new Action Stream. You already do this when the customer abandons a shopping cart. Why not extend the concept when the customer visits your website and looks at low-priced Home merchandise (for instance)?

• Contact 1 = Show customer what the customer looked at.
• Contact 2 = Show customer related products.
• Contact 3 = Your best discount/promotion (since your lust for discounts/promotions requires action).
• Contact 4 = Show customer content options to keep the customer engaged.

In Action Streams, a "-" represents a customer who is about to lapse to a lower / less responsive customer segment.

Remember the segmentation strategy I use to classify customers?

• Elite:  75%+ Annual Rebuy Rate + 12 Month Buyer.
• Loyal:  60% - 74% Annual Rebuy Rate + 12 Month Buyer.
• Quality:  40% - 59% Annual Rebuy Rate + 12 Month Buyer.
• Average:  20% - 39% Annual Rebuy Rate + 12 Month Buyer.
• Struggling:  0% - 19% Annual Rebuy Rate + 12 Month Buyer.
• Lapsed Spend:  12%+ Annual Rebuy Rate (worth spending money on).
• Lapsed Experiment:  5% - 11% Annual Rebuy Rate (worth testing ideas on).
• Lapsed Save:  0% - 4% Annual Rebuy Rate (don't spend money, they are unresponsive).

Any customer with a "-" is a customer in the lowest percentage point of the band. The customer is about to drop down to a lower segment.

• Example:  A Loyal customer drops to a 61% annual rebuy rate. This customer is graded with a "-". This is your signal to DO SOMETHING to prevent the customer from falling from Loyal to Quality status next month.

In your customer database, you flag any customer with a 76% / 61% / 41% / 21% / Lapsed 13% / Lapsed 6% rate ... you flag them as "-", and then you DO SOMETHING to stop the customer from sliding down to a lower customer segment. You set up an Action Stream to mitigate the demise of the customer.

Does that make sense?

Great Moments in Omnichannel History

Traditional retail is in a lonely, unsatisfying alternate timeline.

Action Streams: Compatibility Scores

It's 2001 at Nordstrom.

The e-commerce Executive wants to tailor email campaigns to the merchandise customers previously purchased. Her team tells my team they can support ten (10) different versions of an email campaign twice per week. My team is brilliant. They tell her they can develop a methodology to pair customer purchase history with the content offered in a version of an email campaign.

What they came up with is similar to what I currently call a "Compatibility Score".

Let's oversimplify the process. Say you have two versions of an email campaign, and you want to decide which version the customer should receive. Now, there are better ways to do this, and if you don't like how I oversimplify my example below, by all means, add the level of complexity necessary to suit your needs. But for the love of Nordstrom, do something, ok?!

Ok, here is the merchandise composition of email version #1.

• Mens = 70%.
• Womens = 20%.
• Kids = 10%.
• Home = 0%.
• Accessories = 0%.

Here is the merchandise composition of email version #2.

• Mens = 0%.
  
• Womens = 50%.
• Kids = 30%.
• Home = 20%.
• Accessories = 0%.

Finally, we have a customer with the following weighted purchase history (I weight historical transactions ... often along a 100% for 0-12 month orders, 60% for 13-24 month orders, 35% for 25-36 month orders, 20% for 37-48 month orders, 12% for 49-60 month orders ... your mileage will vary).

• Mens = 15%.
• Womens = 40%.
• Kids = 10%.
• Home = 12%.
• Accessories = 23%.

Which version of the email campaign should the customer receive? I multiply weighted historical percentages by product percentages in each campaign, then sum the multiplied percentages to acquire a "Compatibility Score".

The second version of the email campaign has a 25.4% Compatibility Score, compared to 19.5% in the first version. You'd assign an Action Stream for this customer where the customer receives Version 2. 

Yes, you should test this concept to see if it works for your customers.

As you can see ... it's terribly easy to create a Compatibility Score.

Twenty-four years ago at Nordstrom ... an entire generation ago, we obtained sales increases of 20% within email marketing using Compatibility Scores.

Are you using Compatibility Scores within your Action Streams?

Action Streams and Returns

In case you were wondering, no, "AI" is not going to solve the retail "returns" problem.

It can make your returns problem better, but come on. Most "AI" you see in my industry is Snake Oil, a rebranding of an old problem and an old solution, allowing the salesperson to charge a premium.

Why do I say this?

In 1993 (yeah ... 1993) I built a model using the hyperbolic tangent function (look it up) to predict the return rate of every one of four million twelve-month buyers. Today the snake oil folks would call the hyperbolic tangent function "AI". Anyway, we learned that returns behavior was not predictable until a third purchase. At that point, any customer who returned 60%ish of items (i.e. bought 9 items, returned 6 on those first three purchases) was a customer likely to return in the future, costing us $$$$. Our decision? Cut back on mailings to those customers ... from 20-25 per year to 4-6 per year. Boom. Another million dollars of profit.

In the modern world, some of you have Action Streams for customers who return too much merchandise. There is no reason you have to email customers who return 60%+ of what they purchase. You create an Action Stream once the customer passes a threshold, and you send the customer maybe 4 email campaigns a month instead of 4 per day.

You are under no obligation to keep emailing customers who return most of your merchandise. The customer can still buy from you. But you don't have to waste resources, do you?

Set up an Action Stream to bypass customers who return too much merchandise. You'll also bypass all the snake oil salespeople pushing "AI" solutions upon you.

It's Time!!

Four months go by in the snap of a finger!

Remember just two months ago when you were told you HAD to have a PERECT Cyber Monday experience with 70% off or you'd disappoint today's savvy shopper? Well, how did all that work out for you?

In this run of the Elite Program ($1,000 for existing clients, $1,800 for new clients, no commitment moving forward) I'll look at how many new customers you generated from Cyber Monday and I'll see if they purchased since vs. other customers acquired just before / just after Cyber Monday. That and my usual Rolling Twelve Month Analysis / Comp New Buyer / Comp Segment / Annual Repurchase Metrics / Next Year's Forecast.

It's virtually free. Here's the criteria.

• Pre-pay the fee on or before February 15.
• 5 Years of Customer Data Through January 31.
• Data due by 2/15.
• Analysis completed by 2/28.

Contact me now (kevinh@minethatdata.com) and we'll get busy!

What Would An Action Stream Look Like?

The results were absolutely compelling.

Let's say the customer last purchased eight (8) months ago, and just visited your website. What should you do?

• If you are obsessed with discounts, it's time to kick those into high gear. Instead of 25% off, make it 35% off. You'll love it, your customer will like it.
• Show the customer new arrivals for the product category that the customer previously purchased from.
• Ask the customer if the customer needs to replenish what was previously purchased.
• Promote the items that the customer looked at last month.
• Promote complementary items to the items the customer just looked at.
• Promote complementary items to what the customer purchased previously.
• Ask your customer to engage with more of your content.
• In retail, tell the customer of upcoming in-store events the customer may want to participate in.

When I log into the customer forums on Headphones.com, I am sent into an Action Stream based on the content I looked at.

This stuff works.

And there are dozens of vendors who will help you do it.

So what stops you from doing this stuff?