Free report: “5 Things $39.6M in B2B Programmatic Spend Taught Us.” Three years of managed spend and a billion tracked page loads, turned into five findings platforms won’t put in a pitch deck. Subscribe free to unlock it, plus every new issue of The B2B Stack.
Join 1,600+ B2B marketers, including readers at Dentsu, HubSpot, 6sense, Informa and DirecTV.
Why owning your pipes is the new margin multiplier in the post-DSP era
There was a time when having a seat on a big DSP felt like the pinnacle of digital maturity. Then there was a time when everyone wanted to build their own DSP, whether they were format, channel or niche specialists. Then the market shifted and strategy changed, and everyone wanted the big DSP seats again. If you want the wider context, I covered why AI didn’t kill the dark funnel in an earlier piece.
How AI Rewired the B2B Dark Funnel (and What to Measure Now)
Free report: “5 Things $39.6M in B2B Programmatic Spend Taught Us.” Three years of managed spend and a billion tracked page loads, turned into five findings platforms won’t put in a pitch deck. Subscribe free to unlock it, plus every new issue of The B2B Stack.
And now we are looking at the post DSP era - where the smart play increasingly looks like AI enabled tooling which manipulates containers in the SSP layer, bypassing the weaker signal, high costs and complexity of the enterprise DSP entirely. If you want to see how this plays out in practice, have a look at segment vs signal and running programmatic against the buying committee.
Access to data. Global reach. Precision targeting. Enabled by AI, with a reimagined supply chain that carries less bloat and cost.
The world has shifted. The pipes are now the product.
Today’s smartest B2B marketers aren’t asking how to buy better — they’re asking how to own the system that does the buying.
Welcome to the post-DSP era, where owning your pipes means owning your destiny. For the wider stack, start with the 2026 B2B programmatic guide.
The DSP Was Never Built for B2B
Let’s be honest.
The Trade Desks and DV360s of the world were designed for FMCG at scale — cheap reach, fast frequency, and auction-based liquidity. It gets worse when you layer on firmographic targeting that is stuck on an outdated filing system.
They were never built for:
Smaller audiences that matter more
Account-based segmentation
Multi-signal decisioning
Firmographic precision
Or understanding the real margin dynamics of B2B media, hooking and optimising on composite signals like HVA’s rather then a single conversion event.
That’s why B2B buyers are increasingly paying for complexity they don’t need.
A typical chain looks something like this:
17% DSP fee + layers of extra fees for ad serving, inventory pricing, clearing housing
$1.50 CPM data fee.
10–20%-50% SSP take rate.
And a loss of transparency at every hop.
In practice, you might be paying $1.50 to deliver $0.60 worth of media and you still don’t own the data feedback loop.
The smarter path? Build your own Mini DSP.
Not a Trade Desk clone. In fact, CATEGORICALLY not this.
Instead, think a lean, verticalised buying brain tuned entirely to your business logic, your data, and your economics, keying off the stronger signal coming from the sell side
What a “Mini DSP” Really Means
Forget the idea of racks of servers or engineers writing bidders in C++. Forgot having to ingest millions of random queries per second (QPS), racking up huge 6 figure monthly bills in AWS - I have been there with my PowerLinks business.
A Mini DSP is a composable architecture — a connected set of modular tools that let you run the core functions of a DSP without the baggage, enabled by AI tooling. The identity layer is the hard part, and the CIO you are targeting might not exist in the way your CRM thinks.
The CIO You're Targeting Might Not Exist
Most account-based programmes still run on the same two-step lookup. Step one, build the target account list. Step two, find the CIO or the CTO (or insert other pre-canned job titles) inside each account and point the media at them.
It’s about assembling five layers:
Signal Layer – Where your truth lives.
All the company-level signals you capture across your website, CRM, form fills, and ad interactions. The starting point of your own gravity well.Identity Layer – The map that connects the dots.
Using IP-to-company resolution, hashed emails, or clean-room enrichment to unify fragmented signals into accounts, personas, and buying committees.Decision Layer – The brain.
The logic that decides when and how to bid, what attention is worth, and how to pace toward real business outcomes. This is where AI now plays a huge role.Delivery Layer – The pipes.
Curation layers, SSP containers, or API connections into demand routes like Index Exchange, OpenX, or Bidswitch.
You can even plug into Trade Desk REDS for custom execution if needed — but the key is you control the decision logic before the bid.Feedback Layer – The loop.
Feeding post-view activity, attention metrics, and High-Value Actions (HVAs) back into your system.
The loop is what makes it compound.
Together, these form your Mini DSP — a precision-tuned machine that optimises not for clicks, but for business outcomes. If you are still optimising for clicks, read why clicks are a lie.
The Economics of Control
Once you own the architecture, the economics start to flip.
Instead of paying 17–20% DSP fees, you’re paying for infrastructure, not tolls.
Instead of renting data, you’re feeding your own signal layer.
Instead of being one buyer among thousands, you’re building a private highway between your data and the open web.
The result is more control, cleaner margins, and sharper insights.
Let’s put it bluntly:
The future of programmatic isn’t who has the biggest bidder.
It’s who has the smartest pipes.
The B2B Buying Brain
Here’s what it looks like in motion.
Signals from your analytics platform (Matomo, FunnelFuel Journey, etc.) feed into your data warehouse.
IP resolution connects visits to real companies.
Your JetStream-style platform validates, enriches, and scores those companies based on engagement, fit, and intent.
A decision engine — powered by your own model or an AI Ops Copilot — uses those scores to decide when and where to serve ads.
It bids only on companies that meet your criteria, within curated exchanges you control.
The attention data, post-view engagement, and HVAs loop back into the model.
The system learns.
The pipes get smarter.
Your margin gets cleaner.
This isn’t hypothetical. It’s where the most progressive B2B players are already heading. The signal underneath it is best built by reverse-mapping high-value actions.
Why This Matters Now
Three forces make this moment ripe:
Composability — You can now assemble the same core capabilities as a DSP using APIs and clean rooms.
AI Optimisation — Models can learn pacing, attention, and bid efficiency without needing massive volume.
Signal Ownership — First-party data is the only defensible currency left.
In other words, the tech that once required a 200-person engineering team is now achievable by a focused marketing team with smart partners and a clear spine.
Beyond Media: The Strategic Shift
When you build your own Mini DSP, you’re not just buying differently — you’re thinking differently.
You stop obsessing over CPMs and start optimising for outcomes.
You stop chasing reach and start engineering relevance.
You move from campaign management to signal management.
This is what separates the next wave of B2B marketers from the last.
In the post-DSP era, owning your pipes is the new moat.
Because when you control the flow of data, you control the economics of attention.
The Blueprint: Unlocked for Free Subscribers
Below the fold, I’ll break down:
The exact stack I’d assemble for a real-world Mini DSP
How to source curated supply without DSP middlemen
The clean-room and ID graph vendors that plug together best
The math: how to get your total tech cost under 8% all-in
If you’re serious about building your own buying brain — this is the blueprint.
Below is the full stack, the supply routes, the vendors and the maths, free for subscribers.


