Rebuilding always-on email for Bax Music: who it's for, when it lands, and what's inside.
A twice-a-week email habit ran on autopilot for one of Europe's largest music retailers. We turned it into a strategy — and proved the riskiest part of it with a controlled A/B before rolling anything out.
Client
Every subscriber is a moving target. What someone wanted last month isn't what they want today — and for a retailer with tens of thousands of products and a subscriber base big enough that no two people are shopping for the same thing, a single one-size-fits-all email stops working long before anyone admits it.
Here is how we rebuilt an always-on email program for Bax Music so that every send quietly answers three questions it had never been asked: who is this for, when should it land, and what should be inside it.
A big audience, mailed by habit
Bax Music is one of Europe's largest music-instrument and pro-audio retailers — a catalogue that runs from a first ukulele to a studio-grade mixing desk. Email is a core channel, and Bax was already doing the hard part: two business-as-usual (BAU) campaigns every week, without fail. A Friday-afternoon stock-clearance with six items to move, and a Sunday send of hand-picked products for discovery.
Consistent and reliable — and identical for everyone. Same slot, same logic, whether a subscriber had bought last week or hadn't opened in months. The brief was simple to say and harder to do: keep the reliability, but make every send personalized, better-timed, and worth more — without adding hours to the team's week.
Keep the reliability. Lose the autopilot.
Personalization at scale — on the stack they already run
The goal was personalization: every subscriber getting their own version of Bax's always-on email, not one average message sent to all. We're a senior CRM collective, and we chase that goal a particular way — we get more out of the platform a client already runs, in Bax's case Bloomreach, and coach the team to own it, rather than reaching for new tools. And we measure against a control, so a result is a result and not a story.
The work started, as it always does, with an honest look at what was actually happening. Then it moved across three levers: who each send is for, what goes inside it, and — last of all — when it lands.
A personalized send frequency for every segment
An audience audit is unglamorous and always revealing. We mapped Bax's base by real engagement — from brand-new subscribers, through active openers and clickers, down to the various shades of lapsing and inactive — and found a database contacted by habit rather than behavior. So the first layer of personalization wasn't the content at all. It was the cadence: how often each person hears from Bax now depends on how they engage.
Sorted from newest to least active — each segment on its own rhythm, following email deliverability best practice.
Best practice, made concrete: matching frequency to engagement protects deliverability, cuts wasted sends, and keeps the winnable segments warm — the groundwork every later layer of personalization depends on.
A category signal that learns each customer's taste
This is the heart of the personalization: the content itself. Working from behavior already flowing into Bloomreach, we built a per-customer affinity ranking — using capability Bax already owned — that finds each subscriber's top six product categories.
Reads intent
A purchase counts more than a cart; a cart more than a view. Every signal is weighted by how much intent it truly carries.
Stays current
It leans on recent behavior, so the signal moves as taste moves — this month's interest, not last year's.
Ranks the six
The output is an ordered top six categories per person — the raw material every send now draws from.
Then the real problem: which two, this week?
Defining a top six once is easy. The hard part is deciding which two of the six appear in any given send — for every customer, every week, without a human touching a scheduling sheet. We set up a simple, predictable three-week rotation: two categories per send, three product slots each, cycling so that every subscriber sees all six of their categories over three weeks, then starts again.
Two of a customer's six categories per send — everyone sees all six, then it repeats.
week 1
The read: each week pairs a stronger affinity with a lesser one — 1 + 6, 2 + 4, 3 + 5 — so every send balances a sure-thing category with room to discover. Personalized without being repetitive, automated without being manual; deduplication and quality guardrails keep each send clean, and a sensible default covers brand-new profiles.
All of it runs on the Bloomreach stack Bax already runs. Our job wasn't to add tools — it was to get more out of the one they'd already invested in, and to coach the team to own it. We also closed a real data gap along the way: the cart event wasn't carrying its category, so we reconstructed it and the intent signal finally read complete.
We changed only the content. The audience did the rest.
We never roll out on a hunch. To check whether personalized category content actually earns its place, we split one mid-sized audience straight down the middle — one half got the send as it was, the other got the same email, same time, with the personalization switched on. Opens landed level on both sides, which is the whole point: whatever moved next, moved because of the content.
Read this honestly. Click-through is the result we stand behind fully — measured against a control, opens held level. The revenue figure is real but early, resting on a small pool of first-send buyers, so we treat it as directional — the honest shape of it, not a rounded-up headline.
The month's two best sends came from the same idea
A test is a promise; production is proof. When the signal went live inside the regular BAU calendar, the two sends built on it didn't just do well — they became the two best-performing campaigns of the month. One was a small send, one went out at full scale. Both clicked several times higher than an ordinary always-on email.
The two category-rotation sends versus a typical always-on send.
The read: against a normal-send range of roughly 0.4–1.2%, the two affinity sends topped the month. Modeled at a typical €200–300 order value, that click performance points to roughly €80–120 in revenue per thousand subscribers on the affinity sends — several times an ordinary send. (Illustrative model; the measured revenue figure is the ~€41 → ~€99 above.)
And finally, the timing
With the audience and the content personalized, we turned last to when Bax sends. A brand's default habits are easy to miss: across a full year of business-as-usual sends, engagement wasn't flat across the week. Some days consistently ran strong, while the long-standing Friday-afternoon slot consistently ran weakest — its opens often slipping a day late into the weekend.
A year of always-on sends — shape only: the strong days versus the weakest slot.
Work done: we A/B-tested the calendar and shifted the program toward the days subscribers actually read — away from the weakest slot, without changing a word of the content. The point isn't the exact figures; it's that the timing was tested and tuned, not left on habit.
An always-on program that finally acts like a strategy
- Who — every segment now has its own personalized cadence, matched to how it engages: the winnable groups kept warm, wasted sends cut, and deliverability protected.
- What — the personalized category content is A/B-proven (click-through 1.32% → 1.49%, opens held level), held up as it scaled, and produced the month's two best sends.
- When — the timing was tested against a full year of the brand's own data and tuned away from its weakest slot, toward the days subscribers actually read.
- How — every layer built on the Bloomreach stack Bax already owned, with the team coached to run it, and every claim scoped to what it honestly shows.
Kept honest: click-through is the rigorously controlled result. The revenue figures are directional — real, but drawn from a small first-send base — so we scope them to that send. The audience and timing moves are diagnosed and corrected on the strength of Bax's own year of data, not presented as isolated, separately-tested lifts.
One engagement, three situations it speaks to
Just adopted the platform
You've signed Bloomreach (or Braze, or Adobe) and need it to start producing — fast, and without a rebuild.
Retention revenue has plateaued
Your base is large and your sends are fine, but revenue-per-send has flattened. The lift is hiding in relevance you haven't built yet.
The team is at capacity
You have the ideas and the platform — just not the spare bandwidth to audit, re-time, and personalize on top of business-as-usual.
Most brands don't have a tooling gap
They have a "use what you already pay for" gap. Bax Music's was hiding in who, when, and what. Book a 20-minute look and we'll help you find yours.
Book a 20-minute look