What Salesforce actually paid for
On 20 October, Salesforce agreed to buy Braze for about $5.66 billion, paying $47.50 a share in cash. That is roughly a 58% premium on the previous day's close, for a company that is still loss-making.
The obvious reading is that Salesforce bought a messaging platform. It already owns one. Marketing Cloud has sent more emails than most of us have had hot dinners.
What it actually bought is harder to see on a balance sheet. It bought the moment. Braze's whole architecture exists to notice what a customer is doing right now, in the app or on the site, and respond before that moment passes. Salesforce paid a 58% premium not for a channel, but for timing.
Three moves, one bet
One acquisition is a deal. Three moves in the same week is a pattern.
The same week Salesforce announced Braze, the FTC cleared Publicis to buy LiveRamp for $2.2 billion, on the condition that Publicis keeps LiveRamp's identity layer neutral for everyone else. An agency holding company paying billions to own the plumbing that connects who a customer is across the open web.
Meanwhile, the platforms kept quietly taking levers away from marketers. Meta has removed placement exclusions from yet more campaign types. Google is pushing harder into AI-curated media bundles and automated campaigns, and its new budget allocation controls are guardrails on a machine, not a return of the steering wheel.
A software vendor, an agency group and the two biggest ad platforms are all making the same bet from different directions. The scarce asset in marketing is no longer reach, or creative capacity, or even data in the general sense. It is live context: knowing who this person is and what they are trying to do, at the moment they are trying to do it. Whoever owns that context gets to run the machine. Everyone else gets to feed it.
Why the moment beats the profile
For two decades, marketing technology was built on a slightly flattering assumption: that people are stable. Build a rich enough profile, the theory went, and you would know what someone wanted. Age, income, past purchases, job title, firmographics. A dossier.
Behavioural science has always been suspicious of this. People are not consistent; they are contextual. The same person who agonises over a ₹200 delivery fee will happily pay ₹400 for a coffee on a Friday afternoon. The person who ignores your emails for a year will read every word of one the week their contract comes up for renewal. Preference is not a fixed property of a person. It is something that switches on, briefly, in the right circumstances.
A profile tells you who someone is. A signal tells you what state they are in. And it is the state, not the identity, that predicts what they do next.
That is what the premium on Braze is really pricing. Not a better record of the customer, but a faster read of the customer's moment. The industry is quietly moving from demographics to dynamics.
From pulling levers to reading signals
If you run a marketing team, the uncomfortable implication is that the job you were trained for is being automated from both ends. The platforms are taking away the manual levers: bids, placements, audiences. The data giants are consolidating the context those levers used to act on.
What is left is the part in between, and it is more interesting than it sounds. Three things matter more than they did a year ago.
First, the quality of the signals you feed the machine. An automated campaign optimising toward a lazy conversion event will efficiently find you more of the wrong thing. Choosing what counts as a meaningful signal of intent is now a strategic decision, not a tracking chore.
Second, your own first-party read of intent. If every vendor is selling you the same third-party context, it stops being an advantage. The signals only you can see, in your product, your community, your sales conversations, are the ones a competitor cannot buy.
Third, speed of interpretation. A signal is perishable. The renewal anxiety, the late-night pricing page visit, the sudden spike in a prospect's hiring for a new function: each has a shelf life measured in days. A team that reviews intent data monthly is reading yesterday's newspaper.
The marketer's job is shifting from operating campaigns to deciding which signals deserve to be believed.
What the machine can't buy
There is a counter-current running through the same week's news, and it is worth noticing. While the money flowed into data and automation, the trade press was full of worry about sameness. Marketing Week ran a piece urging marketers to drop AI-speak before the whole industry sounds identical, and another arguing that taste will be what wins in the AI economy.
These two stories are not in tension. They are cause and effect. When every competitor has access to the same context and the same optimisation engine, the machine will converge everyone on the same efficient answer. Efficiency is a commodity the moment it is for sale.
Which means the moment of truth is still won by something no acquisition can deliver. A signal tells you when someone is ready to listen. It does not tell you what to say that they will remember. Timing gets you through the door; distinctiveness is what they recall after you leave. Brands seem to sense this: this month saw Courvoisier launch its first global campaign in twenty years, and a run of big, characterful celebrity-led work.
The winners will be the teams that use the machine for timing and keep the meaning for themselves.
Where signals become insight
The big players have made their bet clear: they will pay extraordinary premiums to know the customer's moment. That should tell every marketer where the value now sits.
But a signal on its own is just noise with good timing. It becomes valuable only when someone interprets it, decides what it means, and responds with something worth hearing. The data layer is being bought up. The interpretation layer is still wide open.
That is the job worth getting good at.
Who is building for business context
Salesforce bought customer context. For most enterprises, the bigger gap is business context: how operations, finances, supply chain and customers connect. That understanding is scattered across systems never designed to talk to each other. A new category of platforms is emerging to close the gap without the multi-year migration that traditional data projects demand.
One example is SCIKIQ, an AI-native data platform that connects the systems a company already runs, governs the data, and adds a semantic layer of business meaning so AI can reason on it where it sits. Forrester has recognised it among the leading AI-augmented data platforms globally, and NASSCOM has named it to its League of 10. Its premise is the Braze deal seen from the other side: what matters is the context, not where the data lives.
Not every company can spend $5.66 billion to buy context. Most of them already own theirs. The work is in making it legible.
Disclosure: the author has worked with SCIKIQ.
