> ## Content Index
> Fetch the complete content index at: https://www.martechtherapy.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Salesforce, Braze and Hightouch: who owns the customer loop now?
- URL: https://www.martechtherapy.com/who-gets-to-own-the-customer-loop/
- Published: 2026-11-02T07:00:00.000Z
- Updated: 2026-11-02T06:59:59.000Z
- Description: Salesforce buys Braze, Hightouch starts sending, and two CDPs pull compute and AI inside your walls. Everyone wants the whole loop now. Here's what that means for ESPs and for you.
- Author: Matthew Niederberger
- Tags: Martech strategy, Customer Engagement Platform, Composable CDP

On 20 October [Salesforce](https://www.salesforce.com/?ref=martechtherapy.com) agreed to buy Braze for $5.66 billion. On 22 October Hightouch announced it sends messages now, across email, SMS, push, in-app and WhatsApp. Its co-CEO, Tejas Manohar, opened [that announcement](https://hightouch.com/blog/lifecycle-studio-customer-engagement?ref=martechtherapy.com) by pointing at the Braze deal.

One company bought an engagement platform.

The other says you don't need one anymore.

[Meiro launched Local AI CDP](http://meiro.io/blog/local-ai-cdp-for-sovereign-ai-inference/?ref=martechtherapy.com) on the same day as the Braze deal, a month after [Zeotap launched Open CDP](https://zeotap.com/news/zeotap-launches-open-cdp-the-first-fully-composable-cdp-that-runs-on-open-source-compute-and-the-customers-own-lakehouse-sovereign-by-design/?ref=martechtherapy.com). I wrote about both [in a recent piece](https://www.martechtherapy.com/sovereign-cdp-open-cdp-local-ai/).

That's a lot of movement for one week, and I wanted a better word for it than an adjective. The best I have is a number.

In the [Martech Market Report](https://www.martechstackbuilder.com/research/martech-market-report-autumn-2026/?ref=martechtherapy.com) I put together for [Martech Stack Builder](https://www.martechstackbuilder.com/?ref=martechtherapy.com), 68 of the 464 tools in the catalog have changed hands since 2023.

That's one in seven.

Twenty deals closed in 2026 before the research cut-off on 11 October, so [Braze](https://www.braze.com/?ref=martechtherapy.com) isn't even in it. Most of the buyers were other software companies, 44 of the 68.

[![One in seven martech tools changed hands since 2023: 8 deals in 2023, 14 in 2024, 25 in 2025, and 20 closed plus 1 agreed in 2026 to 11 October. Buyers: software companies 44, private equity 13, data and information services 5, agency holding groups 4, other 2. Source: The Martech Market Report, Autumn 2026.](https://storage.ghost.io/c/57/c5/57c5ffb5-ecd9-4bf6-b870-68986d981176/content/images/2026/10/who-gets-to-own-the-loop-deals-per-year.png)](https://storage.ghost.io/c/57/c5/57c5ffb5-ecd9-4bf6-b870-68986d981176/content/images/2026/10/who-gets-to-own-the-loop-deals-per-year.svg?ref=martechtherapy.com)

> Everyone is repositioning at once, and they're all heading for the same place: the whole loop of deciding who gets what, sending it, and learning from what happened.

## What did Salesforce, Hightouch, Zeotap and Meiro just do?

Salesforce is buying a modern engagement platform, Hightouch is building one, and Zeotap and Meiro are pulling compute and AI inside the customer's own boundary.

None of it started in October, though.

Here are the last six months:

[![Timeline of martech acquisitions and launches, May to October 2026: Publicis agrees to buy LiveRamp, Databricks launches CustomerLake, MoEngage agrees to buy Aampe, Hightouch launches Lifecycle Studio, Zeotap runs its CDP as a Snowflake Native App, Zeotap launches Open CDP, Hightouch buys Strata, Salesforce agrees to buy Braze, Meiro launches Local AI CDP, Hightouch adds message delivery.](https://storage.ghost.io/c/57/c5/57c5ffb5-ecd9-4bf6-b870-68986d981176/content/images/2026/10/who-gets-to-own-the-loop-six-months-of-repositioning.png)](https://storage.ghost.io/c/57/c5/57c5ffb5-ecd9-4bf6-b870-68986d981176/content/images/2026/10/who-gets-to-own-the-loop-six-months-of-repositioning.svg?ref=martechtherapy.com)

**Salesforce is buying the loop.** The offer is $47.50 a share in cash, about 58% above Braze's last close but still below the $65 it listed at in 2021, and it's expected to close in Salesforce's fiscal 2028\. Braze brings real-time event-driven engagement, deep mobile and in-app, and the AI decisioning it picked up when it bought OfferFit in 2025\. All of that is supposed to strengthen Marketing Cloud, which is now called Agentforce Marketing, yes, it got renamed again.

Salesforce has owned an email platform since ExactTarget in 2013, and has spent the last couple of years rebuilding it on Data Cloud (also renamed, Data 360). Buying a direct competitor in the middle of that rebuild tells you how the rebuild is going. I mean, in the past twelve months three of my own clients left Salesforce Marketing Cloud for a CEP, and I'm one consultant in Delft.

[Braze - Martech Stack BuilderCross-channel customer engagement platform for orchestrating personalized messaging across push, email, SMS, in-app, and web in real time.![](https://static.ghost.org/v5.0.0/images/link-icon.svg)Martech Stack BuilderMartech Stack Builder![](https://storage.ghost.io/c/57/c5/57c5ffb5-ecd9-4bf6-b870-68986d981176/content/images/thumbnail/braze-280d04499cff-5f34cda7-c1a6-402b-874e-ca5ef4ab1381.png)](https://www.martechstackbuilder.com/vendors/braze/?ref=martechtherapy.com)

**Hightouch is building the loop.** It started as reverse ETL, the thing that synced your warehouse into Braze and every other engagement platform. Then came identity resolution, audiences, AI decisioning, Lifecycle Studio in June, and on 9 October the acquisition of Strata, an integration startup founded by three early Iterable employees. Now it's delivery, with domain warming, consent controls and its own deliverability team. Medium says it moved its sending over in under a month and runs millions of messages a day through it.

Manohar says they've been building delivery "for a couple years," which I believe. The product is older than the Braze news. The timing of the announcement isn't.

[Hightouch - Martech Stack BuilderComposable CDP that activates data warehouse models as audiences, syncing to 200+ marketing, sales, and advertising destinations without data duplication.![](https://static.ghost.org/v5.0.0/images/link-icon.svg)Martech Stack BuilderMartech Stack Builder![](https://storage.ghost.io/c/57/c5/57c5ffb5-ecd9-4bf6-b870-68986d981176/content/images/thumbnail/hightouch-849b3a77c4be-142dce97-488f-4478-bf94-b327a5de1b9a.png)](https://www.martechstackbuilder.com/vendors/hightouch/?ref=martechtherapy.com)

**Zeotap and Meiro are claiming the layers underneath.** Zeotap took its compute off the proprietary warehouse and onto open-source Spark in the customer's own storage. Meiro put the AI model inside the deployment, so customer data doesn't travel to someone else's model to be reasoned about. Different end of the stack, same instinct: own a layer that somebody else used to charge you for.

[![Concept diagram of which layer of the martech stack each vendor now claims. Salesforce, through Braze, claims message delivery and decisioning; Hightouch built message delivery; Zeotap Open CDP added compute and storage on open-source Spark; Meiro Local AI CDP added the AI model; a standalone ESP claimed nothing new.](https://storage.ghost.io/c/57/c5/57c5ffb5-ecd9-4bf6-b870-68986d981176/content/images/2026/10/who-gets-to-own-the-loop-who-owns-which-layer-1.png)](https://storage.ghost.io/c/57/c5/57c5ffb5-ecd9-4bf6-b870-68986d981176/content/images/2026/10/who-gets-to-own-the-loop-who-owns-which-layer-1.svg?ref=martechtherapy.com)

## Is the composable CDP turning into a suite?

In practice, yes. With delivery added, Hightouch's composable CDP now covers most of what a CEP does.

In June I wrote that Hightouch had rebuilt, piece by piece, the packaged CDP it once told buyers not to buy. Event collection, identity, audiences, analytics. Delivery does the same thing to the CEP.

[The CDP you weren't supposed to buyHightouch said the packaged CDP was dead weight. Four years on, Gartner lists it in the CDP Magic Quadrant. The architecture was right. The pitch skipped a step.![](https://storage.ghost.io/c/57/c5/57c5ffb5-ecd9-4bf6-b870-68986d981176/content/images/icon/Martech-Therapy-Icon-Only-Square-aa22e33e-7ca0-441c-8380-2fde99ce8777.png)Martech Therapy: CDP strategy and agentic AIMatthew Niederberger![](https://storage.ghost.io/c/57/c5/57c5ffb5-ecd9-4bf6-b870-68986d981176/content/images/thumbnail/The-CDP-you-werent-supposed-to-buy-2ab61607-8db5-49a5-84ab-41173339453d.png)](https://www.martechtherapy.com/the-cdp-you-werent-supposed-to-buy/)

Where the composable argument went last time ⬆️

The announcement says it in two sentences that sit a few paragraphs apart. The first: "Paired with the Composable CDP, that means enterprises no longer have any need for an engagement platform in the traditional sense." The second: "We believe that the most composable architecture will win."

To be fair, both can be true.

Hightouch will still send through your existing CEP if you're not ready to move, and I don't doubt it means that. But composable now means you're **allowed** to keep your other tools. It doesn't mean you're encouraged to.

And honestly, the logic holds up.

In June, Braze, [Bloomreach](https://www.martechstackbuilder.com/vendors/bloomreach/?ref=martechtherapy.com) and [Iterable](https://www.martechstackbuilder.com/vendors/iterable/?ref=martechtherapy.com) all announced [Databricks](https://www.martechstackbuilder.com/vendors/databricks/?ref=martechtherapy.com) tie-ups that handed the data layer to the warehouse and kept execution for themselves. Four months later, a warehouse-native company is coming for the execution.

It's also the opposite of what I liked about [Databricks CustomerLake](https://www.martechtherapy.com/databricks-customerlake-the-warehouse-cdp/). Databricks built the data, the decisioning and the workflow, then handed the send to the engagement platforms, which I took as a lesson learned from Twilio buying Segment and ending up in competition with its own partners. Hightouch has chosen to fight the channel instead of feeding it. That might work. Twilio's version didn't.

[Did CEPs just hand the data layer to Databricks?Three engagement platforms made the same bet in the same week. Whether it pays off is more contingent than the announcements suggest.![](https://storage.ghost.io/c/57/c5/57c5ffb5-ecd9-4bf6-b870-68986d981176/content/images/icon/Martech-Therapy-Icon-Only-Square-6ca11950-f046-4889-976f-f1256f8759a7.png)Martech Therapy: CDP strategy and agentic AIMatthew Niederberger![](https://storage.ghost.io/c/57/c5/57c5ffb5-ecd9-4bf6-b870-68986d981176/content/images/thumbnail/Did-CEPs-just-hand-the-data-layer-to-Databricks-1-5ba70bc2-0166-473f-83eb-dc0b6367d8dd.png)](https://www.martechtherapy.com/did-ceps-just-hand-the-data-layer-to-databricks/)

The June deal that set this up ⬆️

## Why is martech consolidating now?

I see two reasons.

AI is making the application layer cheaper to build, which puts pressure on valuations, and AI decisioning learns best from the outcomes of messages you send yourself.

The money first.

Software stocks have been sold off since late 2025 on the fear that agents and AI-built tools will replace a lot of what companies pay SaaS vendors for. By May, Salesforce and Adobe were trading at roughly half their early-2025 earnings multiples. Klaviyo dropped 32% in a day after its Q1 results, while growing 28%. Braze was down to about 60% of its end-of-2024 price in September, before Salesforce's offer lifted it back above where it started.

[![Martech share prices indexed to 100 at the end of December 2024. By 26 October 2026: Twilio 244, software ETF (IGV) 111, Braze 109, Salesforce 74, Adobe 58, Klaviyo 50, HubSpot 38. Source: Yahoo Finance monthly closing prices.](https://storage.ghost.io/c/57/c5/57c5ffb5-ecd9-4bf6-b870-68986d981176/content/images/2026/10/who-gets-to-own-the-loop-martech-stocks-since-2025.png)](https://storage.ghost.io/c/57/c5/57c5ffb5-ecd9-4bf6-b870-68986d981176/content/images/2026/10/who-gets-to-own-the-loop-martech-stocks-since-2025.svg?ref=martechtherapy.com)

Hightouch's announcement makes the same argument from the other side: "The whole 'workflow layer' of Martech is being eaten by AI." It quotes Juan Mendoza's [survey of enterprise martech teams](https://themartechweekly.com/enterprise-martech-outlook/?ref=martechtherapy.com), where lifecycle marketers "have literally stopped using their engagement platform in any form I recognize."

If the middle of the stack gets cheaper to build, the obvious move is to own more of the ends. Each player does it from a different position, though. Salesforce is a public incumbent defending its multiple, so it buys growth and a working AI product. Hightouch is private, and my guess, or rather my reading of how funding rounds tend to work, is that a bigger share of each customer's invoice makes a better story for the next one. Zeotap and Meiro are independents squeezed by Databricks and Snowflake moving up into CDP territory, so they claim the layers a warehouse can't easily copy.

Now the second reason. AI decisioning learns from outcomes: delivered, opened, clicked, converted, unsubscribed. When someone else sends the message, those signals come back late and partial, through webhooks and exports. When you send it yourself, you own the feedback your model learns from. Hightouch doesn't say this anywhere, so take it as my interpretation, but I think it explains the move better than the invoice does.

I made a version of this point in [the Agentic CDP series](https://www.martechtherapy.com/your-agentic-cdp-will-optimize-what-you-measure/). An agent optimizes whatever you measure. Whoever sends the message decides what gets measured.

## What does this mean for ESPs and CEPs?

Sending infrastructure and ESPs with their own customer base should be fine. The horizontal campaign suites in the middle are the most exposed.

An ESP does two jobs. It gets the message into the inbox, and it gives you everything around that: the campaign builder, segments, templates, reporting. Those two jobs are heading in different directions.

Getting it into the inbox is *turning into infrastructure*. [Amazon SES](https://www.martechstackbuilder.com/vendors/amazon-ses/?ref=martechtherapy.com), [SendGrid](https://www.martechstackbuilder.com/vendors/sendgrid/?ref=martechtherapy.com) and Postmark already sell it that way, and every platform that adds sending needs one of them underneath. They're probably better off this month than last.

The stock market seems to agree: Twilio, which owns SendGrid, is worth more than twice what it was at the start of 2025, while the suites lost between a quarter and almost two thirds of their value. [Twilio also owns Segment](https://www.martechstackbuilder.com/vendors/segment/?ref=martechtherapy.com) and sells far more than email, so that's a hint, not proof.

ESPs that own a customer base or a vertical are protected by distribution rather than features, at least as a business. [Klaviyo](https://www.martechstackbuilder.com/vendors/klaviyo/?ref=martechtherapy.com) and its Shopify merchants are the obvious example. Its revenue is still growing; its share price halved anyway. Investors are pricing the category, not the company.

The ones I'd worry about are the horizontal campaign suites in the middle. Too heavy to be infrastructure, too thin to hold the decision. If the decisioning platform picks the audience, writes the content and triggers the send, the campaign builder has nothing left to do. The risk for them isn't one dramatic migration. It's orchestration moving out first, after which the ESP turns into an interchangeable line item at the next renewal.

We need to be honest in stating that running an ESP is hard. It is.

IP warming, inbox placement, the Gmail and Yahoo sender rules, SMS carrier registration, all of it is slow specialist work, and it gets harder the more senders share the infrastructure. Hightouch says it has the expertise in-house, and Medium is a credible reference. Doing it for years across hundreds of brands is a different test.

How that goes?

Honestly, no idea yet, and neither does anyone outside Hightouch.

Hightouch also hasn't said publicly what the sending runs on. For a buyer that matters, because it decides who owns your sender reputation.

## What should buyers ask before their next renewal?

Mostly, where your lock-in moved to, because every one of these foundations is rented from somebody.

All four moves are sold as stronger foundations. Go one layer further down, though, and each of them sits on something it doesn't own. Zeotap's open compute runs in somebody's cloud, and a model you can host inside your own deployment is almost always an open-weight one that somebody else trained. Hightouch's sending runs on infrastructure it hasn't named yet. And Braze's customers will be renewing with Salesforce in a year or so, whether they ever chose Salesforce or not.

None of that is a criticism, since every stack is rented at some level. What it means is that lock-in doesn't go away when a company repositions. It moves to whichever layer that company now owns.

Scott Brinker's [newsletter on the Braze deal](https://newsletter.chiefmartec.com/p/martech-consolidation-infrastructure-and-the-braze-salesforce-deal?ref=martechtherapy.com) has a subtitle that says the same thing from the buyer's side:

> "Everybody wants consolidation until the rent comes due."

So before your next renewal, I'd ask a different set of questions than "packaged or composable."

Whose foundation am I tied to now, and did that change this month?

If I leave, who keeps the sending domains, the IP reputation and the consent history? Where does the outcome data go, and in whose format?

And who on my team actually understands the layer that just changed hands?

That last one is a readiness question more than a technology one, and in my experience it's the one most teams struggle with.

[The Martech contracts dates and rights87% of Martech contracts renew themselves and 86% say nothing about whether your customer data trains someone else's model. The paper decides.![](https://storage.ghost.io/c/57/c5/57c5ffb5-ecd9-4bf6-b870-68986d981176/content/images/icon/Martech-Therapy-Icon-Only-Square-2a22f496-12c3-47d6-a0c2-ec382e8fdbec.png)Martech Therapy: CDP strategy and agentic AIMatthew Niederberger![](https://storage.ghost.io/c/57/c5/57c5ffb5-ecd9-4bf6-b870-68986d981176/content/images/thumbnail/The-Martech-contracts-dates-and-rights-11b9fcfa-07fd-4a5f-ac47-ac71b2215e8a.png)](https://www.martechtherapy.com/the-martech-contracts-dates-and-rights/)

What to check in the contract before the renewal conversation ⬆️

The loop is what everyone is selling now. Make sure you know which parts of yours you actually own.

Do you have any questions after reading this article?  
Or need support with your Martech projects?

[Contact me today >> ](https://www.linkedin.com/in/matthewniederberger/?ref=martechtherapy.com)