John Wanamaker's hundred-year-old line — "half the money I spend on advertising is wasted; the trouble is I don't know which half" — has survived every marketing revolution since. Here's the uncomfortable update: with modern attribution, you can know which half. Most teams don't, because their reporting stack was never designed to answer the question. This is how to build attribution that holds up — and use it to find the 20–30% of budget you can reallocate with confidence.
Why Last-Click Reporting Keeps You Blind
Most B2B teams still run on last-click attribution dressed up as insight. The problem isn't just that it over-credits the final touch — it's that it systematically misleads reallocation decisions:
- It rewards closers, not openers. Branded search and "contact us" direct visits hoover up credit for demand that paid media, content and events actually created. Cut what looks weak under last-click and you cut the top of the funnel first.
- It punishes anything with a long lag. In B2B, 90+ day cycles are normal. Channels that create demand early look like they "don't convert" for two quarters — long enough for a nervous budget review to kill them.
- It can't see offline. Sales conversations, word of mouth, webinar follow-up calls — all invisible, all credited elsewhere by default.
The result is a predictable pattern: teams over-invest in bottom-funnel capture, under-invest in demand creation, and then wonder why pipeline thins out six months later. The blind spots behind this dynamic are mapped in detail in attribution blind spots costing you 30% of budget.
The 3-Layer Model That Makes Reallocation Safe
Single-model attribution fails because different questions need different lenses. The practical answer is a 3-layer model:
- First-touch (demand gen view). Which channels create net-new demand? Use for budgeting awareness and acquisition spend. Judge campaigns on qualified first touches, not closed revenue.
- Multi-touch (nurture view). Which touchpoints move deals forward between first contact and opportunity? Use weighted multi-touch (even U-shaped is fine) to compare nurture channels — content, email, retargeting — on influence, not conversion.
- Self-reported attribution (validation layer). Ask every new lead one field: "How did you hear about us?" Then compare against what your tracking says. Where they diverge, your tracking model is wrong — and self-reported data is usually closer to the truth for offline and dark-funnel sources.
Layering matters because reallocation decisions get safer with each layer: first-touch tells you where demand comes from, multi-touch tells you what accelerates it, and self-reported keeps the whole system honest. The full build-out of this framework is in the 3-layer attribution model that works.
The Minimum Tracking Setup (Before You Trust Any Number)
No model survives bad inputs. Before acting on attribution, the plumbing has to be consistent:
- Disciplined UTM convention. One documented schema — source, medium, campaign naming — enforced everywhere, including sales-generated links. One wildcard UTM in a spreadsheet corrupts a quarter of reporting.
- CRM source fields that don't get overwritten. Original source, latest source, and campaign-of-record fields with write-once logic. If SDRs can hand-edit source data, they will.
- Closed-loop revenue reporting. Every opportunity linked back to first touch and influencing touches. No revenue connection, no ROI attribution — just channel traffic reports.
- Server-side or robust pixel tracking. Browser-only tracking loses 20–40% of conversions to blockers in 2026. If your numbers silently shrink month over month, suspect the pipe before the channel. What a modern, resilient setup looks like is covered in from pixel to purchase: tracking that actually works.
This is deliberately unglamorous work. It's also the difference between attribution you defend in a budget meeting and attribution you defend in an argument.
Finding Your 30%: The Reallocation Audit
With the model and tracking in place, run a structured audit across every active channel and campaign:
Step 1 — Rank by incremental contribution, not total. For each channel, compare results against a holdout or a paused period if you have one. Channels that only "perform" under last-click go to the review pile first.
Step 2 — Separate capture from creation. Branded search, direct, "contact us" pages are capture — they harvest demand someone else made. Fund capture just enough to keep conversion rates healthy; the marginal pound belongs to creation.
Step 3 — Apply the 3-layer test. A channel earns its budget if it passes at least one layer convincingly: strong first-touch volume, clear multi-touch influence, or frequent self-reported mentions. Channels that fail all three are your reallocation candidates.
Step 4 — Reallocate in tranches, not cuts. Move 20–30% of the failing spend into proven creators over two months, watching pipeline creation rates. Tranching protects you from being wrong once — and from the organisational panic of a visible budget cut.
Teams that run this audit with honest attribution typically free up 20–30% of budget — not by discovering magic channels, but by stopping quiet waste: the display retargeting with a frequency cap problem, the event sponsorship nobody attributes leads to, the "brand awareness" spend with no first-touch data at all.
Making It Stick: Attribution as a Monthly Ritual
A one-off audit decays fast. The teams that keep the 30% advantage treat attribution as governance: a monthly review where reallocation decisions are made against the 3-layer model, logged, and checked against last month's predictions. The cadence, agenda and decision rules for exactly that process are laid out in budget reallocation: the monthly governance ritual — treat it as the operating system this article installs.
The compounding effect is the real prize. Reallocating even 5% of spend from dead channels to working ones every month doesn't just save budget — it steadily raises the average performance of everything you spend.