An emerging marketing category

Marketing knows who.
It still guesses what to say.

Every channel, audience, and timing signal is optimized. The actual message — the words a customer sees — is still written in advance and hoped for the best. Message Decisioning closes that gap.

$200B spent on direct marketing globally every year
<1% of messages are truly personalized at the individual level
20–50% lift from message-level optimization

Messaging is the last thing
marketing hasn't optimized.

The industry built sophisticated systems for targeting, timing, and delivery. Then wrote the message weeks in advance and called it personalization.

01

Static journeys can't adapt

Customer journeys are locked at launch. When behavior changes between Monday and Friday, the message doesn't know. It fires anyway.

02

Rules written years ago

Hand-coded if/then logic made sense at the scale of 2019. Today it creates irrelevant experiences for millions — and no one has time to rewrite it.

03

A/B tests that finish too late

By the time your experiment reaches statistical significance, the campaign moment has passed and customer context has shifted entirely.

04

Messages pre-decided before reality happens

Copy is written in a campaign brief — weeks before a customer does anything. The message has no idea what the customer did this morning.

What is Message Decisioning?

Message Decisioning is the decision layer between customer data and message delivery — the system that selects the right message for each individual, at the moment of interaction, and learns from every outcome.

It doesn't write copy. It doesn't manage channels. It doesn't run campaigns.
It decides which message should be sent to this specific person, right now — and gets smarter every time.

The unit of analysis is the individual — not the segment

Segmentation assigns people to groups. Message Decisioning operates on each person directly — reading their specific combination of signals, history, context, and behavior at the moment of interaction. Two people in the same segment, with the same product, at the same stage of their journey can receive completely different messages. Not because they belong to different sub-segments. Because they are different people.

Not this

A/B Testing

Picks a winner from two options, once, slowly. Message Decisioning selects from many options, for each individual, continuously.

Not this

Personalization Tokens

Merges a first name or segment into a template. Message Decisioning selects the message structure, tone, and content itself.

Not this

Campaign Automation

Executes pre-written messages at scheduled triggers. Message Decisioning decides what to execute at the moment of execution.

Not this

Generative AI Copywriting

Creates new message content on demand. Message Decisioning selects among available messages — it decides which to send, not what to write.

Not this

Journey Orchestration

Sequences the steps and timing of a customer journey. Message Decisioning determines the content of each step — not the order of steps.

This

Individual-level selection

Every customer gets the message best matched to their context, behavior, and history — from the available message inventory.

This

Continuous optimization

Each outcome — a click, a conversion, an ignore — teaches the system. The next decision is smarter than the last.

This

Stack-agnostic layer

Integrates between your data and your delivery channels. The missing layer — not a replacement for what you've already built.

The layer marketing
was always missing.

Every other layer of the marketing stack was productized. One wasn't.

01
Data Layer
CDPs, data warehouses, behavioral signals
Existing
02
Audience & Targeting
Segmentation, lookalikes, suppression logic
Existing
03
Message Decisioning
What to say to each individual, decided in real time
Missing
04
Delivery & Automation
ESPs, push platforms, SMS, in-app channels
Existing
05
Analytics & Measurement
Attribution, reporting, campaign dashboards
Existing

It complements your stack. It doesn't replace it.

Message Decisioning plugs in between your customer data and your delivery infrastructure. Your existing tools keep doing what they do well.

What changes: messages are no longer written in advance and hoped for the best. They're decided at the moment each customer is reached — from a library of available options, optimized for that individual's context.

Key distinction

Other tools execute messages.
Message Decisioning decides which message to execute.

The Decision Loop

Every interaction follows a continuous loop. No frozen logic. No one-time experiments. Each cycle feeds the next.

Step 01

Decide

Evaluate the individual's context — behavior, history, timing, constraints — and select the best available message from the library.

Step 02

Deliver

Pass the selected message to your existing channel infrastructure. Email, push, SMS, in-app, web — no new pipes needed.

Step 03

Observe

Record the outcome — click, conversion, ignore, unsubscribe. Every signal matters, including silence.

Step 04

Learn

Update the model. The next decision for this individual — and others like them — is informed by what just happened.

This only became possible recently.

Message Decisioning isn't a new idea. It's a newly feasible one. Four forces converged to make it practical at enterprise scale.

01

Data Density

Enterprises now have years of historical data linking specific message elements to individual-level outcomes. The training data now exists.

02

Model Capability

AI can evaluate message-individual fit at the scale of millions of simultaneous decisions. Individual-level inference in real time is no longer computationally exotic.

03

Economic Pressure

Rising customer acquisition costs mean the cost of the wrong message has never been higher. Message-level optimization now has a clear and measurable ROI.

04

Operational Feasibility

Message selection can now be automated without armies of analysts. What previously required weeks of A/B testing now runs continuously in the background.

Small lifts. Enormous impact.
Because campaigns repeat.

20–30%

typical lift in response rates from message-level optimization

25–35%

improvement in conversion rates when messaging matches individual context

↓ CAC

each improvement reduces marginal acquisition cost across every future campaign

Enterprise marketing is a repeating system, not a single event.

Most enterprise programs run the same campaigns — acquisition, cross-sell, retention, reactivation — month after month, across millions of customers. That repetition is the key insight.

A lift in message response rate is modest in isolation. Compounded across 12 campaign cycles, across multiple product lines, across a customer base of millions — it becomes a material financial outcome. Message Decisioning doesn't just optimize a campaign. It improves a system that runs indefinitely.

The compounding principle: A message sent to 2 million customers, 10 times a year, with a 1% conversion rate improvement generates hundreds of thousands of additional conversions — without spending another dollar on media.

The Segmentation Value Gap

Most marketing programs treat message selection as a segmentation problem: divide the audience into buckets, assign a message to each bucket, test periodically. This approach feels rigorous. It leaves enormous value on the table.

Segmentation optimizes the average message for a group. But within every segment — no matter how precisely defined — individuals vary dramatically in what motivates them. A "high-value urban millennial" segment still contains people who respond to aspiration, people who respond to proof, people who respond to peer validation, and people who respond to simplicity. The segment gets one message. Most of them get the wrong one.

Message Decisioning operates at the individual level — below the segment floor. It doesn't replace targeting strategy. It captures the value that targeting leaves behind.

Built for regulated, complex environments.

Decision automation in enterprise marketing must coexist with regulatory, brand, and operational controls. Message Decisioning is designed with this reality built in.

Eligibility & Compliance Rules

Hard constraints — regulatory, legal, or policy-based — are enforced upstream of the decision layer. The system only selects from permissible options for each individual.

Frequency & Channel Controls

Communication limits, suppression lists, and channel preferences are respected. Optimization happens within defined guardrails, not around them.

Auditability & Explainability

Every decision can be traced. Why was this message sent to this individual? The system produces a clear record — important for brand, legal, and analytics teams.

Policy-Safe Modeling

Models are trained and evaluated against brand guidelines and regulatory constraints. Optimization targets outcomes, not just engagement metrics that can mislead.

Workflow Integration

Message Decisioning embeds directly into existing campaign execution systems. It doesn't require parallel infrastructure — it plugs into what you already operate.

Secure Data Handling

Customer-level signals are processed within enterprise data governance frameworks. No raw customer data needs to leave the environment for decisioning to function.

What it looks like in the real world.

The difference isn't theoretical. Here's how Message Decisioning changes outcomes for real marketing teams.

The system isn't assigning people to message buckets. It's reading each person's specific combination of signals — purchase history, browsing behavior, prior message responses, life context, communication style — and selecting the message that fits that individual. The same person can be price-sensitive and aspirational and recently anxious about a billing issue. The message reflects all of it.
Financial Services — Credit Card Acquisition

The same offer. The message that makes it real is different for every person.

A bank prospects for a rewards card. Every qualified prospect sees the same product. What varies is every single message they receive.

Before "Earn 3X points on every purchase." One message, written for an imagined average. Sent to everyone. The traveler, the deal-hunter, the cautious first-timer, the parent stretching a budget — all receive the same words. Most don't respond. The system doesn't know why.
With Message Decisioning One prospect has frequent flight searches, a recent hotel booking, and opened two prior travel-themed emails. Her message leads with where the points take her — not the rate. Another has comparison-shopped three cards this week and responded to a cashback offer last year. His message leads with proof: ranked against alternatives, with numbers. A third has no credit history and searched "how to build credit." Her message is calm and clear — what this card does, how approval works, no pressure. A fourth opened the last two emails but didn't click, and has a young child on the account. His message connects the card to family spending patterns — the everyday utility, not the aspirational use. None of these people were put in a segment. The system read each one individually and selected accordingly.
Financial Services — Personal Loan Acquisition

Same rate. What converts each person is specific to them.

A lender targets pre-qualified prospects. The product is identical for all of them. The message that moves each one isn't.

Before "Get up to $25,000. Rates as low as 7.9%. Apply in minutes." The rate-and-amount frame — designed to be universally relevant, which means it's rarely specifically compelling. Low conversion despite a competitive offer.
With Message Decisioning One prospect has three open credit card balances and searched debt consolidation twice this month. Her message is about relief — one payment, one rate, one less thing to manage. She's not being told about loan features; she's being shown a way out of a specific feeling. Another prospect recently searched home renovation contractors and has a mortgage. His message is about possibility — what he could do, framed as an investment in the place he already chose. A third opened the email, read to the bottom, and didn't apply — twice. His message this time leads with transparency: the exact monthly payment for the amount he'd likely qualify for, no surprises, no pressure. The rate is identical for all three. What makes it real to each of them is entirely different.
Telecom — Retention

Churn prevention that doesn't train the whole base to expect a discount.

A carrier flags at-risk accounts. The instinct is to offer a discount. The cost of doing that to everyone is substantial — and it teaches customers that leaving is how you get a better deal.

Before "Stay and save $20/month." Sent to everyone flagged as at-risk. Works for some. For the rest, it either doesn't move them — or it successfully trains them to churn-and-return next cycle for another offer.
With Message Decisioning One customer has been on the network for nine years, has never called support, and her churn signal is just reduced app usage. She doesn't need a discount — she needs acknowledgment. Her message is about her tenure: what it means, what it's worth, a quiet expression of appreciation with no ask attached. Another customer called support twice last month about a coverage gap and got transferred twice. His churn isn't about price — it's about a specific, unresolved frustration. His message addresses that directly: the issue, what changed, and what he can do if it happens again. A third customer has been price-comparing plans and responded to a discount offer two years ago. She is price-sensitive. She gets the offer — but only her. The system doesn't need a rule that says "discount sensitive customers." It reads the pattern and acts on it individually.
Insurance — Cross-sell

The next product introduced through each person's actual frame of reference.

An insurer cross-sells home insurance to existing auto policyholders. Same product, same underwriting, same price range. What makes someone ready to hear about it depends entirely on who they are right now.

Before "Bundle your home and auto and save up to 15%." Written for the average customer. Sent to all eligible policyholders at 60 days post-auto-renewal. Resonant for almost no one specifically — and timed to the policy calendar, not to the customer's life.
With Message Decisioning One customer closed on a home 47 days ago. She's in the thick of ownership — new locks, a basement that concerned her, a neighborhood she's still getting to know. Her message is about protection, not savings. Another customer has called twice to understand his coverage limits. He's a careful reader of policy documents and responded positively to detailed communications. His message is about completeness — one provider, one point of contact, full picture. A third customer is seven months into a significant home renovation and has an auto claim from last year. His message acknowledges the claim positively — handled well — and connects that experience to what comprehensive coverage would mean for a home under active renovation. None of this is demographic targeting. It's reading the actual signals each person has generated and selecting the message that meets them there.

Message Decisioning across industries.

Click any case to expand ↓
Non-Profit — Donor Acquisition

Finding first-time donors before they've given anywhere

+

A nonprofit runs acquisition campaigns to prospect lists — lapsed event attendees, peer referrals, affiliated community members. Conversion from prospect to first-time donor is low.

Before "Your gift changes lives." Mission statement, impact photo, donation ask. The same appeal sent to everyone — designed to offend no one, and therefore specific to no one. The person who attended the gala, the one who was referred by a friend, and the one who googled the cause at midnight all receive the same words.
With Message Decisioning One prospect attended the gala two years ago, engaged with two social posts about community impact, and has given to a local cause before. Her message is about belonging — she's already part of this, this is just the next step. Another prospect clicked through from a data-driven news article about program outcomes, spent four minutes on the impact report page, and has never donated anywhere. His message leads with numbers: what $50 funds specifically, how outcomes are measured, third-party validation. A third prospect was referred by a close friend who's a major donor. She's curious but skeptical — her browsing pattern shows she read the overhead ratio page twice. Her message doesn't ask her to give. It shows her what giving looks like from the inside: who decides, how it's spent, what accountability looks like. The mission is the same. The message is built for each person's specific combination of motivations and hesitations.
Non-Profit — Donor Retention

Keeping donors giving — year after year

+

A nonprofit's biggest revenue challenge isn't acquisition — it's lapse. A large percentage of first-year donors never give again. Renewal campaigns underperform not because the mission changed, but because the message doesn't remember who gave and why.

Before Annual appeal letter with year-end urgency. "Renew your support before December 31st." The same message to a lapsed first-year donor, a seven-year monthly giver, and someone who gave once after a disaster — all of whom gave for completely different reasons, and all of whom need a completely different reason to give again.
With Message Decisioning One donor gave her first gift after watching a video about a specific family's story. She hasn't given since. Her renewal message reconnects to that story — what happened next, told specifically. Another donor has given monthly for six years, missed two months, and opened the last three emails without acting. His message acknowledges the relationship directly: what his cumulative giving has meant, not a generic annual summary. A third donor gave $500 after a specific crisis event two years ago and has never given outside that context. Her message isn't a renewal — it's an update on that exact situation, written as if the organization remembers why she gave in the first place. Because it does.
Financial Services — Collections

Recovering payments without destroying relationships

+

A lender sends collections communications across early, mid, and late delinquency. Repayment rates are low. The wrong tone at the wrong moment — for this specific person — damages the relationship and the brand.

Before Escalating urgency templates by delinquency bucket. "Your account is past due. Act now to avoid further consequences." Compliance-drafted, legally safe, humanly inert. The customer who missed a payment for the first time in nine years receives the same letter as the one who has missed six in a row.
With Message Decisioning One customer has a nine-year payment history with no prior misses, called the support line last week, and has a joint account with a recent large hospital charge. Her message leads with acknowledgment — this is clearly unusual for her — and offers a clear, low-friction path to resolution without implying she's a credit risk. Another customer missed payments twice before, responded to a structured payment plan offer both times, and has been on a plan successfully. His message leads with the plan option, framed around what worked for him previously. A third is six weeks delinquent, has multiple accounts showing stress, and has never responded to a prior communication. His message is direct, clear, and unambiguous — no false warmth, no buried call to action. The regulatory floor is identical for all three. Everything above it — tone, structure, what the message leads with — is specific to each person's actual history and situation.
DTC Subscriptions — Wine, Pet Food & More

Converting trial customers into loyal subscribers

+

A subscription brand acquires trial customers through promotions. The challenge: converting a discounted trial into a full-price, long-term subscriber. Churn after the first box is high — and indiscriminate discount offers make it worse.

Before "Your subscription renews in 7 days." Followed by a discount offer to prevent cancellation. Works for a small slice. For the rest, it either doesn't move them — or it works, but only by training them to cancel-and-rejoin whenever they want a deal.
With Message Decisioning One customer rated three wines, spent time on the tasting notes page, and added two bottles to a wishlist. Her message is about what the curation is becoming — her specific palate profile, what the next box will reflect, the sense that someone is paying attention. Another customer ordered the trial, received it, and has had no further engagement. He's not disengaged — his open rate is fine — but nothing has connected yet. His message is practical: what the subscription actually costs per bottle, how it compares to what he'd otherwise buy, the flexibility to skip. A third customer ordered the trial as a gift, mentioned it in a post-purchase survey, and lives in a city with a strong gifting culture. Her message reframes the subscription entirely — not as something she maintains, but as something she gives. None of these people received a discount. The system identified who needed one — and who didn't — by reading their actual behavior, not by applying a rule.
Online Gaming — Monetization & Retention

Turning free players into paying ones — without burning them out

+

A mobile or online game uses push notifications and in-game messages to drive purchases, re-engagement, and season pass upgrades. Blanket promotional messages are increasingly ignored — or lead to uninstalls.

Before "Don't miss the limited-time event! Get 2X gold this weekend only." Urgency-and-scarcity push sent to every player. High opt-out rate among casual players. Competitive players already knew. Lapsed players feel like they're being sold to, not welcomed back.
With Message Decisioning One player logs in daily, completes every ranked match, and has checked the leaderboard 14 times this week without purchasing. His message is about exactly where he stands — his rank, the gap to the next tier, what it would unlock. It doesn't mention a sale. Another player plays primarily when her friends are online, has purchased twice after friend-activity notifications, and hasn't logged in for three weeks. Her message is about what her friends are doing right now — not an event, not a discount, a social pull. A third player has spent 40 hours in the lore sections of the game and has never made a purchase. His message introduces new story content that's genuinely relevant to the arc he's been following — framed as continuation, not upsell. A fourth player purchased the season pass last year, didn't this year, and has reduced play sessions since the new season launched. Her message addresses the gap directly: what's different this season from last, in the specific areas where she previously spent her time.
Streaming & Media — Subscription Retention

Keeping subscribers who are drifting, not yet gone

+

A streaming service identifies subscribers showing low engagement. Churn risk is rising but they haven't cancelled yet. The window to re-engage them is narrow — and a generic content push often closes it.

Before "New this month on [Platform]." Generic content newsletter featuring the 3 titles marketing chose to promote. Sent to all low-engagement subscribers regardless of what they watched, when they drifted, or why they joined in the first place.
With Message Decisioning One subscriber watched all six seasons of a legal drama, finished it three months ago, and hasn't started anything since. Her message isn't a top-10 list — it's one specific show, chosen because it shares the same writers, a similar moral complexity, and a cast member she's seen in two other things she rated highly. It arrives with enough context to feel like a recommendation from someone who was paying attention. Another subscriber has two young kids, watches in 20-minute sessions, and hasn't opened the app since his commute routine changed. His message isn't about content at all — it's about the mobile download feature, specifically for offline viewing on a commute. A third subscriber joined for a specific documentary series, finished it, and has browsed but never committed to anything else. Her message surfaces the next documentary the platform is releasing in her area of interest — not as a promotion, but as a heads-up. Each message is built from what that specific person actually did, not from a re-engagement template.

Message Decisioning is emerging.
Follow its development.

Frameworks, case studies, and analysis on the discipline of message-level intelligence. No product pitches.