Kevin Hillstrom: MineThatData

Here we go, as promised! There are many questions. Skip the ones you don't want answers to.

Steven: In all your meetings, whether working for or consulting with a brand, was there ever a moment when you saw the participants grasp what you were imparting on them? It would be great to hear an example of how you saw a team or group work, either from the ground up or top down to really change the culture and business trajectory.

It is rare for an entire group of people to grasp a concept at the same time or on a similar timeline. The core themes of my work (customer acquisition > customer retention ... incremental response vs. reported response ... importance of new merchandise and winning items) are not concepts that are accepted by average professionals. Imagine being the person who manages paid social and I tell the person that 70% of the conversions he observes in his reporting via Facebook would happen anyway if he didn't advertise?

There are individuals who grasp concepts. Those individuals generally run with the idea. I work with a CEO/Owner (for close to fifteen years) - this guy gets it and in general does a fabulous job implementing the concepts. When I analyze his data, I can see when he hires somebody who disagrees with him ... customer data changes and his business doesn't grow as fast. This would be an example of a top-down approach. The leader asks staff to execute a certain way, and sometimes the team working for him chooses otherwise.

I worked with a Marketing Vice President who took the concepts and ran with them on his own, independent of the rest of the business. Merchandise productivity was bad, his marketing investments in new customers were good, the business grew as a consequence. When the individual left, the new marketing person identified "low hanging fruit" ... cutting marketing spend significantly. The business crumbled, and has not recovered.

I am currently working with a team that is getting results ... the individuals may or may not agree with me, but they are working together to make change happen. Their Leader sets the tone, the team respects the Leader, consequently, they listen to me. As you'd expect, implementation is bumpy ... it always is ... and it doesn't matter ... the team knows they want to see better performance and they are working together to make it happen. People have to be able to get along with each other for this to happen.

Anne: Increasingly, I am seeing nonprofits I work with say they are not going to do postal mail because of cost. They want to use email and social only. They invite me to be on their board but they do not want to listen. I want to give them guidance but I do not want to lead them astray because conditions are changing rapidly and nonprofit budgets are thin and endangered. Any help would be appreciated.

Ok print lovers, here is your moment ... I'm about to land on your side! There are three issues with print ... costs are increasing, response is declining, and those who respond are generally 65+ years old. For non-profits, do you care if those who respond are generally more than 65 years old?

If costs increase, response has to increase in response to increased costs. This is the challenge - nonprofits I've worked with don't know if response increases or decreases ... they sometimes see marketing costs as a budget line, but they don't marry response to costs.

My catalog clients have something they call the "profit factor" ... it's the percentage of sales that flow through to profit. Let's pretend that factor is 40%. If a catalog costs $1.00 to put in the mail, then the client has to generate $1.00 / 0.40 = $2.50 of sales in order for the segment to break-even. If the cost to put the catalog in the mail increases to $1.10, the client has to generate $1.10 / 0.40 = $2.75 of sales in order for the segment to break-even. My clients would cut circulation on anybody expected to generate $2.50 to $2.74 to offset the increased cost of the mailing.

You should be in a similar circumstance ... your mailings must offset the cost of the mailer and generate enough donations to cover overhead and have enough left over to actually do good work. Which means you probably need to generate between $3.00 and $5.00 per recipient in donations to offset the cost of direct mail. And if direct mail costs increase by 10%, you need to generate $3.30 to $5.50 to offset increased costs.

As long as you generate enough donations to offset costs and have plenty left over to do good work, print is important regardless of cost increases.

Liz: I understand your advice on building communities outside of Meta and Google. What is the expected range of outcomes from these programs? How long does it take to generate results via new customers, and what can I expect in terms of share of new customers from building communities?

Here's an answer my readers won't like ... it will likely take 3-4 years of hard work to make a difference. Also ... you have no choice but to do the hard work.

This isn't a place where you want a direct marketer leading the efforts. At all. The direct marketer doesn't have great ideas in this realm, and will quit long before the program has a chance of succeeding.

This isn't the place where you want an ROI-based digital marketer leading the efforts. At all. This individual will also quit long before the program has a chance of succeeding, citing low "ROAS". The individual would rather pay Facebook money today for orders tonight.

I know of a company that spent six years building the community. In years 4/5/6 efforts paid off. Roughly half of their sales come from new customers, they believe roughly a quarter of their new customers come from their community of registered users. They believe that each registered user is worth $10 net sales per year. The reality is that these are guesses, but they aren't bad guesses based on the data shared with me.

The community must be passionate about what the company sells (is anybody passionate about what Macy's or Kohl's sells, for instance?) and feel like there are benefits to being in your community ... purchasing your merchandise ... over what somebody else sells.

The first year will be a hot mess ... virtually no positive results ... a lot of numbers wonks asking you why you are wasting so much time/money on something so pointless?

The concept of a "registered user" is important - you have to track if your community is visiting your website and/or looking at your merchandise. If you can demonstrate that a registered user visits your website monthly, you're on to something.

Robby: How is AI influencing prospecting and what is the best way to take advantage of it? Is any company excelling at it?

In digital marketing, AI has already stamped itself over everything you do. It determines the content/merchandise in email messages sent to non-buyers. Ask the Listraks and Klaviyos of the world for examples. Search is already aligned with AI ... when I search for HEADPHONES WITH A NEUTRAL SOUND SIGNATURE Google isn't looking for keywords, Google is inferring my meaning via AI.

AI-centric marketplaces will likely supplant modern search in the next 5-10 years, assuming the AI wonks don't boil all of our fresh water cooling their servers.

I have a direct mail client who uses AI models for reactivation, and has good results ... good meaning +20% over their baseline models. This is a bare-bones use of AI, of course ... they're just using AI like one would use a regression model.

Most of my clients tell me that classic direct mail prospecting is "dead", and when something is "dead", AI isn't going to be of much help. I recently saw a mail plan comparing 2026 to 2016 ... prospecting circulation was down 80% due to a combination of terrible response and increased costs. AI cannot overcome structural changes in an industry, which is why we don't see many AI solutions applied to direct mail paired with outstanding outcomes ... if that existed, the print folks on LinkedIn would be screaming from the mountaintops - their glee would be unavoidable.

Honestly ... if 40% of an email subscriber list has not purchased anything historically, I'd gather whatever attributes I could for this audience and use AI to align messaging and merchandise with non-buyers.

Bob: What advice of yours changes when a new business or a business with sales < $1,000,000 a year has a handful of unique products/SKUs where product/market fit is still evolving, and where the database is < 10,000 customers?

The general theme of my advice doesn't change, the level of complexity required to implement ideas changes considerably. If the small business wants to grow, it needs to find clever ways to identify likely prospects - and the small business will have to do it "on the cheap". The organic social route is very important for these companies. It's very important until the business exceeds $10,000,000ish per year ... then the rules begin to change and there needs to be a more-generous marketing budget.

The small business needs to understand if the product they sell leads to repeat purchases. If the product is a one-and-done product, well, it's all about finding new customers for a long time. If the small business is selling body wash? Well, that's a different business model, and a small number of very happy customers can do a lot of free marketing on behalf of the brand.

Small databases change the way experimentation happens. If you have 10,000 customers and 5,000 have an email address, you are likely to just split the file in half when experimenting with email marketing concepts. The small brand is likely to segment customers into a small number of segments so that any analysis/results are statistically relevant.

The small business that has one or two key items needs to closely monitor when those items are dying. The slightest miss against budget/expectations should set off alarm bells. There needs to be a plan for the small business ... an exit plan for a limited assortment (i.e. be bought by a larger brand or become an Amazon supplier) or a product development plan to become a "brand".

The small business looks a lot less at customer metrics and looks a lot more at conversion metrics. Where is traffic coming from? How does traffic convert? Do customers buy multiple items or a single item (this becomes an important segmentation variable)? Are customers clicking-through email campaigns? Analytics sophistication grows as net sales grow. It doesn't mean the small brand isn't analytically driven ... quite the opposite ... but the small brand worries much less about complex questions/solutions, focusing instead on what it controls and what it can understand with limited resources.

One area where the small business is no different from a large brand is in purchase activity within the first 3-4 months following a first purchase. Regardless of ecommerce brand, it is really important to convert a customer to a second purchase quickly before the customer fades away. Small businesses that focus on this metric have an inherent advantage over other small businesses.

Using a "simple is better" approach, the small brand is infinitely more nimble than Macy's ... an enormous advantage.

Ed: We used to mail a 184 page catalog, but shifted to 124 pages when paper / printing / postage costs increased. Around that time our sales slumped. Did sales slump because we featured fewer pages in catalogs?

Believe it or not, I first analyzed this challenge in (checks notes) 1991. THIRTY-FIVE YEARS AGO! We had an equation at Lands' End that calculated the elasticity between pages offered and merchandise sold. Each additional page (i.e. going from 160 pages to 164 pages) generated incremental sales, but at an ever-decreasing rate.

At Eddie Bauer in (checks notes) 1999 ... TWENTY-SEVEN YEARS AGO we had an equation that optimized circ depth, page count, and quality of product offered. Small page counts were super-charged with best-selling products ... allowing 96 pages to perform nearly as well as 248 pages from a sales standpoint (meaning we could mail 96 pages forever because of the profit advantage those pages had).

But times change.

If you are like many modern catalogers ... you have an approximate 75% organic percentage ... page counts are nearly irrelevant. You heard that correctly. Your customer has moved on, and responds digitally. Spend your time on social, search, email, video, community ... it's over.

If you are like many B2B brands ... you have an approximate 35% organic percentage ... it's like it is the mid-90s. Your catalog IS THE REASON why the customer purchases. For these customers, you have to show the customer your merchandise assortment or the customer might not look for those items online.

Be willing to experiment ... if you mail 124 pages ten times per year, go ahead and mail 184 pages one time per year, and reduce circulation by a third in response. Try it. See what happens. Also be willing to test ... conduct a mail/holdout test and see if your organic percentage (the share of sales that still happen if a catalog is not mailed) is 35% or 75%.

Also - be willing to check your merchandise productivity. Most of the time pages offered is minimally relevant while merchandise productivity is maximally relevant. If customers like what you sell less year-over-year, it's easy to misattribute that issue to something less meaningful (like email marketing click-through rates and/or catalog pages and/or traffic from Instagram).

        

A quarter century ago, our Nordstrom stores were designed to have Cosmetics near the in-mall entry to our store. Nothing like having fragrance constantly pumped into the mall to attract attention.

Of course, having Cosmetics near the in-mall entry door meant we attracted a lot of new customers ... Cosmetics customers.

And that was a good thing! Why? Because our analytics indicated that a new Cosmetics buyer would quickly cross-shop into other categories ... especially Women's Footwear, located right next to Cosmetics. From there the customer willingly cross-shopped into Women's Apparel and Accessories. Now we had a four-category customer ... our analytics indicated that a customer buying four times ... once from each of four categories ... was more valuable than a four-time buyer within just one category.

In other words, the success of the customer was pre-built into the design of the store.

What is pre-built into the success of the Macy's ecommerce customer?

A generation later and we've abandoned the concept of pre-building customer success into our designs.

For instance, here's a popular Shopify store.

Surely somebody at Kith understands the merchandise that brings in high-value new customers.

Right?

Surely you understand the "entry points" into your ecommerce store that generate high-value new customers.

Right?

        

Send me your questions (kevinh@minethatdata.com) and I'll dedicate a post to answering the questions y'all submit.

Think of it like free consulting - except all of our readers benefit.

Fire away! I'll accept questions through the end of day 3/11/2026.

        

In a recent query, I used a logistic regression equation to measure multiple-division buyers. Are these buyers more likely to purchase again compared to other buyers?

• 0-12 Month One Time Buyers = +27% more likely to purchase again.
• 0-12 Month Two Time Buyers = +10% more likely to purchase again.
• 0-12 Month 3x+ Buyers = +3% more likely to purchase again.

        

This might come as a surprise to some of you, but your customers don't universally love everything you sell.

Every customer has a favorite merchandise division/department/category. This "favorite" division/department/category dictates what you should show the customer in the future.

Example: A Women's Apparel customer (favorite historical division) will spend 48% of future demand on Women's Apparel (your mileage will vary). The second-highest division generates just 12% of future demand, the third-highest division generates just 8% of future demand. This customer should see Women's Apparel, obviously. The marketer needs to be disciplined with this customer.

Example: A Home customer (favorite historical division) will spend 8% of future demand on Home merchandise, with three divisions/departments/categories showing significantly higher future shares of merchandise purchased. Guess what? This customer needs to be cross-shopped. The marketer can be creative with this customer.

Each customer essentially warrants a customer-based merchandise assortment.

Do you think about your customers in this manner?