Your next best
customer already
ordered. You didn't
notice.

A few years ago a professional athlete placed an order with a brand I was running. Nothing about it stood out. Same shipping label, same middling cart total, same confirmation email as the two hundred orders around it. I only caught it because the name registered while I was scrolling the admin looking for something else entirely.

I sent a handwritten note out with the package. He wrote back, and the reply is the part I still think about: he had bought it because most of his team already had one. A locker room of customers we had never counted, one conversation away, and the only reason I found out is that I happened to glance at a name on a Tuesday.

That is a nice story. It is also a terrible system. Because the real question it raised was not how lucky I got. It was how many of those had already come and gone while I was busy being a competent operator doing everything else.

Every order in
Shopify looks exactly
the same.

Open your orders view. Every row carries the same four things: a name, a total, a date, a fulfillment status. Nothing in that interface separates the person who runs merchandising at a national retailer from the person who found you through a Pinterest board at midnight. Shopify is not doing anything wrong here. It is an order system, and orders are what it shows you.

The gap is that the identity behind an order is some of the most valuable information your store collects, and it arrives with the value already stripped out. You get an email address, a shipping city, and a cart. Everything that makes that customer interesting lives outside your admin, in the public record, and nothing joins the two together.

This matters more now than it did five years ago, for reasons that are not very inspiring. In 2025, paid ad CPMs rose 8.64% year over year (Triple Whale, Ecommerce Benchmarks 2025), so the same budget buys less attention than it did the year before. Retail shelf space is harder to win than it was. The creators actually worth knowing are buried under sponsorship pitches they never asked for.

A customer who already bought your product and liked it skips every one of those queues at once. They are the cheapest warm introduction available in commerce, and most stores hand them back.

Five people are
buying from you,
and one is famous.

When people hear VIP they picture a celebrity. That is the least useful version of this idea, and chasing it is how brands end up with a gifting program and no revenue. Across the brands I have operated and advised, the people genuinely worth catching in an order feed sort into five types, and the celebrity is the one that matters least.

That last row is the one brands underrate, and it is the only one you can already solve with data you own. As of 2026, around 73% of customers never come back for a second order (Shopify, Ecommerce Customer Acquisition, July 2026). The ones who do come back deserve a different grade of attention than the ones who do not.

That concentration has been visible for a decade. Back in 2015, RJMetrics found that the top 1% of ecommerce customers were 18 times more valuable than the average customer (RJMetrics, 2015 Ecommerce Buyer Behavior Benchmark, May 2015). A decade of platform changes later, the shape of that curve is the same one I see in the brands I operate and advise. A top customer is sitting in your Shopify reports right now, and most stores treat their twelfth order exactly like a stranger's first. If you fix nothing else in this post, fix that. While you are in there, check your definition of a top customer against how brands miscalculate customer lifetime value. The definition is usually where this goes wrong.

The wholesale row is the one with the clearest dollars attached. A retail buyer ordering a single unit through your DTC site is running due diligence, and that order is the earliest signal you will ever get of a wholesale conversation. Most brands find out months later, through a cold email from the same person, and negotiate from a weaker position. If you have not been through one yet, a DTC brand's first wholesale account covers the margin maths waiting on the other side.

The manual playbook,
which genuinely
works.

None of this needs software to start. Here is the version I ran by hand for years, in the order I ran it.

What to check, in order

Start with the email domain. A corporate domain on a direct-to-consumer order is the highest-signal field you already have and it costs nothing to notice. Someone ordering from a retailer's domain, a magazine's domain, a fund's domain, or a competitor's domain is telling you something the cart total never will. This one field catches most of the wholesale and press opportunities on its own.

Then the name paired with the shipping city. A name alone is noise, because names repeat constantly and confidence built on a name alone is how you end up emailing a stranger about their fictional career. A name that agrees with a city in a public biography is a real signal. This pairing is less about finding matches than about refusing bad ones, which turns out to be the more important job.

Then order value against your own baseline, not against somebody's benchmark. A cart at four or five times your average deserves thirty seconds of human attention regardless of who placed it. Sometimes it is a wholesale buyer. Sometimes it is a customer about to become your best one.

The weekly habit

Block twenty minutes on a Monday. Sort last week's orders by value and scan the top twenty. Scan every order from a non-consumer email domain. Anything that catches your eye gets thirty seconds of searching, and anything that survives those thirty seconds gets a line in a shared doc with the follow-up written next to it.

The doc matters more than it sounds like it should. The failure mode here is almost never missing the customer. It is finding the customer, feeling briefly clever about it, and then never following up because the day filled in. Write the play down beside the name or the play does not happen.

Then your order
volume quietly ends
the habit.

Here is the arithmetic that kills it. At fifty orders a week, twenty minutes covers your entire feed comfortably. At five hundred you are sampling, while telling yourself you are covering. By the time you are at five thousand you have stopped altogether, and the honest version is that the habit died about three months before anybody noticed.

There is a second failure that is subtler and worse, which is that the window closes. A creator who ordered six weeks ago and heard nothing has already formed an opinion of your brand, and it is a neutral one. Press moves faster than that. A buyer moves slower, but spends the gap talking to somebody else in your category. Detection that arrives that late is just trivia.

The third problem is the one that makes people wary of automating any of this: false positives. Names collide, constantly. If a system tells you someone notable just ordered and it turns out to be a different person with the same name, you will send an embarrassing email. After the second time, you will stop trusting the system entirely. Any version of this, manual or automated, lives or dies on corroboration rather than on detection.

So I built the
thing I wanted
to exist.

I built VIP Alert because I wanted the Monday habit to run without me, and because nothing I could find did the job honestly. It reads who each Shopify customer actually is at checkout. Then it grades how much that person matters to your specific store, and writes the answer back into Shopify as native customer tags and order notes.

The write-back is the part I care about most, and it is the part that took the longest to get right. Because the answer lands as ordinary Shopify tags, anything that already builds audiences from tags picks it up with no extra configuration. Klaviyo, Postscript, Shopify Segments and Flow all just see it. That means a detection can trigger a flow you have already built rather than becoming another dashboard nobody opens. If your flows are not in that shape yet, build them first. The Klaviyo flows every Shopify brand should have live is the sequence I would start with, before adding any of this on top.

The interesting engineering problem was not finding people. It was refusing to. Most of the real work went into corroboration. Matching a hometown against a shipping city, checking employer signals, requiring agreement across sources, and confirming the person is living before anyone gets an alert. In our own test store it identified an Olympic snowboarder by matching his hometown to the shipping city, and correctly declined to flag a deceased athlete who happened to share his name. The second half of that sentence is the actual product.

The thing I got wrong early was assuming every store wants the same VIPs. They do not, and it is obvious in hindsight. A wholesale-forward label wants retail buyers and is largely indifferent to creators. A beauty brand wants precisely the opposite. So the grading is weighted per store rather than universal. That is the kind of correction you only get by putting the thing in front of merchants who tell you it is wrong.

Building an app again after several years of advising app founders has been a useful humbling. Everything I say to founders about scope, onboarding and free tiers is considerably easier to say than to do, which is roughly what I expected and still worth relearning firsthand. If you are weighing whether to build something yourself, building a Shopify app in 2026 is the reality check I would want you to read first.

The alert is
worthless without
the play.

An alert with nobody attached to it is a notification, and notifications get muted. Decide three things before you turn any detection on, manual or automated: who acts, how quickly, and what exactly they send. Write it down once and it holds for a year.

None of those rows contain an ask, and that is deliberate. The instinct when a creator orders is to reach immediately for a partnership, and it is the wrong first move nearly every time. Being the brand that noticed, said thank you, and wanted nothing converts far better six weeks later than the brand that opened with a proposal. The structured influencer marketing funnel comes after that, and it works better when the first contact was not a pitch.

The recommendation itself still moves product. In 2026, Salsify found that 23% of shoppers had bought something because a social media influencer recommended it (Salsify, 2026 Consumer Research Report). A creator who bought your product with their own money is already past the part most brands pay for.

Measure it the boring way. Tag the customers you actually actioned, then compare their repeat rate and lifetime value against the rest of your file ninety days later. If VIP handling is doing anything, that cohort separates visibly. If it is not separating, the problem is what you are sending rather than who you are finding. Use DTC repeat purchase and retention benchmarks as the baseline so you are comparing against something real instead of against a feeling.

Somewhere in your last thousand orders is a buyer, a creator, or somebody who writes about your category for a living. You already paid to acquire them. The only open question is whether you find out while there is still something useful to do about it.

Questions brands ask
about VIP customers.