01
S2 · Retention programs
Lifecycle & retention programs — the operating system under your customer contact
The core engagement. Contact strategy, LTV tiering, churn prediction, orchestration across every channel — designed and then built.
02
Short answer
A lifecycle and retention program decides who gets contacted, when, on what evidence, and what the business expects each contact to be worth. I design that system and then build it with your team inside your own stack — the models, the segmentation and the orchestration — so the thing that goes live is the thing that was specified.
What gets built
A risk score tells you who is at risk. A survival curve tells you when — and only the second one can have an intervention scheduled against it. That distinction is the shape of the whole program: decide on evidence, then act at the moment the evidence points to.
I design the program, then build the models and automations with your team, in your stack. Four things come out of it, in this order, because each one depends on the last.
A contact strategy with a budget
Not a channel plan — a contact budget. How many times a given customer can hear from you in a month before the marginal message costs you more than it earns, and which messages get the slots. Most programs have never priced this, which is why the calendar wins every argument.
LTV tiering that survives contact with finance
Lifetime value modelled per cohort and per acquisition channel, reconciled with how your finance team already defines contribution. A CRM team with its own private definition of LTV loses every budget conversation it enters.
Churn prediction you can act on
Survival models — usually gradient-boosted accelerated failure time — that give you a time-to-event, which is the only version an intervention can be scheduled against. Built in your environment, on your data, with your team watching how it is done. The case, across four markets at Tibber.
Orchestration across the channels you actually have
One decision layer that decides what each customer should receive next, sitting above five channel teams each optimising their own open rate. This is where most of the measurable gain sits, and it is almost always an integration problem wearing a creative costume.
- Scope
- Contact strategy · LTV tiering · churn prediction · orchestration
- Model
- Built with your team, inside your stack
- Duration
- A quarter and up, scoped per phase
- Recent
- Vivino · Schibsted · Compricer
- Stack
- Braze · Salesforce Marketing Cloud & Data 360 · Databricks · SQL · Python
03
How a program runs
Phased, so the commitment is a phase at a time.
01
Diagnose
Two weeks in your data. Where retention leaks, what it’s worth, and which of the four pieces above is actually the constraint. Ends in a scope with a number on it — or an honest recommendation that you don’t need this.
02
Design
The contact strategy and the target state, argued in front of the people who’ll have to live with it. Written as a specification your engineers can read, not as a slide.
03
Build
Models, segmentation, flows, QA. With your team, pairing where they want to learn it, and documented as we go.
04
Prove it
Holdout groups from day one. Read the numbers together at the end of the quarter and kill what didn’t work — including my ideas.
If churn is the number under pressure, send the shape of the problem.hello@andersson.consulting
04
Where the judgement comes from
The scale behind the design decisions.
- 30+
- marketsCRM run across IKEA’s European organisation — governance, localisation, one stack, many countries.
- 20+
- brandsMulti-brand lifecycle orchestration at Bauer Media.
- 550k
- customersFour countries, as Head of CRM & Lifecycle Strategy at Tibber.
- 23
- teamsReviewed and scored for AI leverage — an internal program inside one organisation.
In-house leadership
- IKEA
- Adobe
- Bauer Media
- Tibber
Independent engagements
- Schibsted
- Compricer
- Vivino
We already have a CRM agency. Does this replace them?
Usually not. Agencies are generally good at production and weak at economics, because that isn’t what they’re paid for. I work on the layer above — the strategy, the models, the decision logic — and your agency keeps producing against a brief that now has a reason behind it.
Do we need a data team for this to work?
You need customer-level data somewhere queryable and someone who can grant access to it. You don’t need a data science team, because the modelling is the part I do hands-on. If the data itself is the mess, sorting that out becomes the first phase and I’ll say so during diagnosis.
How do you measure whether the program worked?
Holdout groups from day one, agreed with you before anything ships, and read against retention and contribution rather than opens and clicks. If a change can’t be measured against a holdout, that’s a reason to be suspicious of it — including when it’s mine.
How long before we see something?
Something in production inside the first quarter, always. The structural work — the tiering, the survival models, the orchestration layer — takes longer, and I’d rather be honest about that than promise a quarter and deliver a pilot.
What does it cost?
Scoped per phase, so you commit one phase at a time and the Diagnose phase is deliberately small. Send the shape of the problem and what comes back is a scope and a number for the first phase. The pricing page sets out how each phase is quoted.
06
Related services
S1
Fractional Lifecycle Leadership
The retention seat, held from inside your organisation — including the person who writes the queries.
Ongoing, sized to the role
S3
The AI Marketing Audit
Every team and workflow reviewed and scored for where AI creates leverage — and where it’s theatre.
Three weeks, fixed price
07
Available for new engagements
Start a conversation
If the timing fits, email me directly. There is no funnel behind it, and the reply comes from me.