Why adding channels stopped working
In 2019 you could win by simply being present. Launch a Meta campaign, send a newsletter, print a stand for the trade show, and growth followed attention. In 2026 attention is cheap to buy and expensive to keep. Acquisition costs across European B2B and DTC have risen sharply since 2020, cookie-based attribution no longer carries the weight it used to, and buyers touch six or seven surfaces before they ever fill in a form.
The companies still growing profitably are not the ones running the most channels. They are the ones whose channels share a single brief, a single data layer and a single follow-up engine. What follows is the structure we use when we take over a growth programme: five layers, each with one job, in the order they actually pay off.
Layer 1 — Foundation: a brand that holds, and data that tells the truth
Nothing above this layer compounds without it.
Brand strategy and identity decides what your traffic means. Two companies can buy the same keyword at the same price and convert at a threefold difference, because one answers “why you” in four seconds and the other asks the visitor to work it out. Positioning, naming, a messaging hierarchy and a visual system that survives contact with a 320-pixel screen — that is infrastructure, not decoration.
Analytics and insights decides what your spend means. Before any budget scales you need reliable server-side tracking, a definition of a qualified lead that sales actually agrees with, channel-level attribution, and cohort reporting that shows what a customer is worth at month twelve rather than on day one. Most of the waste we find in audits is not bad creative. It is good creative judged on the wrong number.
Rule of thumb: if you cannot say what a lead from each channel is worth to within a small margin, you are not ready to double any budget.
Layer 2 — Demand capture: performance marketing that pays for itself
Capture is the cheapest demand you will ever buy, because the intent already exists. Search, retargeting, high-intent LinkedIn and marketplace placements all belong here.
The work that makes this layer profitable is unglamorous and systematic:
- A creative testing cadence — a fixed number of new concepts every month, not every quarter, with a written kill rule.
- Offer testing kept separate from creative testing, so you know which variable moved the number.
- Landing paths built per campaign angle, instead of one page carrying every promise.
- Bid and budget decisions tied to the cohort value from Layer 1, not to cost per click.
Capture has a ceiling, though. Only so many people are searching for what you sell this month. Once you own that demand, pushing more spend into the same layer simply raises your cost per acquisition. That is the moment companies usually blame the agency. It is actually the signal to open Layer 3.
Layer 3 — Demand creation: PR, events and offline media, worldwide
This is the layer most growing companies skip, and the reason their capture layer plateaus. Demand creation makes people want the category before they go looking for it.
PR and communications buys credibility that no auction sells. Media relations, founder positioning, commentary in the trade press your buyers actually read, and a crisis plan written before you need it. In regulated and enterprise markets, one well-placed piece in a sector publication outperforms months of cold outreach.
Events and experiential remains the highest-converting channel in B2B — conferences, launches, trade-show stands, roadshows. The difference between a stand that costs money and a stand that makes money is almost entirely what happens afterwards: a scanned badge with no sequence behind it is a business card in a drawer.
Traditional media and OOH — print, billboards, transit, radio, TV, sponsorships — is not nostalgia. Offline is measurable now: geo-lift tests, dedicated landing paths, branded-search lift and incrementality windows. Planned worldwide and read against the same dashboard as paid, OOH often turns out to be the cheapest branded-search driver a company owns.
The purpose of this layer is not reach. It is to make Layer 2 cheaper.
Layer 4 — Compounding: AI automation and lifecycle engines
Layers 2 and 3 generate volume. This layer stops that volume from leaking.
AI automation systems take the operational work that scales linearly with growth and make it scale at close to zero marginal cost: lead enrichment and routing, quote and proposal generation, first-line support, research and reporting agents, internal copilots over your own documents. The test for a good automation candidate is simple — it happens often, it follows rules, and a human currently does it from a template.
Lifecycle and email engines are where margin lives. Onboarding sequences, activation nudges, retention and win-back, and transactional messages that do a job rather than confirm receipt. Owned channels cost nothing per send and convert at rates paid media cannot reach, yet they are usually the last thing a growing company builds. Reverse that order.
Built together, these two change your unit economics rather than your traffic.
Layer 5 — Ownership: engineer what you shouldn’t rent
Eventually the system outgrows off-the-shelf tools: the dashboard your team actually needs, the portal your clients keep asking for, the internal tool that replaces nine spreadsheets, the product your service business has been giving away for free.
SaaS engineering at this stage is a growth decision, not an IT one. Owning the layer where your data and workflow live means your automations stop breaking when a vendor changes an API — and it is often how a service company turns delivery into a product with real margin.
Sequencing: a realistic first ninety days
- Weeks 1–3 — measure. Tracking, lead definition, attribution, baseline cohorts. Do not change spend yet.
- Weeks 3–6 — sharpen. Positioning and messaging, then rebuild the two landing paths carrying most of your traffic.
- Weeks 5–10 — capture. Restructure paid around intent and start the creative cadence.
- Weeks 7–12 — compound. Ship the lifecycle sequences and the first two automations.
- Weeks 10–14 — create demand. First PR placements, event calendar locked, one offline test with a measurement design attached.
Three months is enough to fix measurement, capture and follow-up. Demand creation is a six-to-twelve-month instrument. Both should start now — they simply report on different clocks.
Four mistakes we see most often
- Scaling spend before measurement. Every euro you add amplifies a number you don’t trust.
- Judging channels in isolation. OOH and PR lift branded search; scoring them on last-click guarantees you will switch off the thing that was working.
- Automating a broken process. Automation makes a bad workflow fail faster. Fix the process first, then take the hands off it.
- Treating owned channels as an afterthought. Email and lifecycle are the only part of the stack whose cost does not rise with scale.
Where to start
You don’t need a strategy deck to find out which layer is holding you back. Start with the free assessment closest to your suspicion — each takes a few minutes, scores you out of 100, and ends with a prioritised roadmap plus a report in your inbox.
And if you would rather talk it through: we work with companies across Europe and run campaigns worldwide, digital and traditional, under one measurement framework.