Most B2B paid channel mixes are a polite guess: last year's split, nudged, with no model underneath and nothing to show finance when they ask. Every budget change means rebuilding the same fragile spreadsheet from zero. The fix is a planning model built on real benchmarks, along with your account history, that projects the return per channel, per budget, before you spend.
Planning is a different job from attribution: attribution scores the past, planning decides the future, and you often need planning first for a lot less. Harry is that model, a semi-automated planner that turns your annual or periodic budget into a channel mix and a model you can defend, before a penny goes out the door.
Below we’ll show you how to turn paid campaign budget allocation from a guess into a number you can actually defend.
Open your current paid media channel budget plan and ask yourself one honest question: where did these numbers actually come from?
For most teams, the real answer is last year's split, slightly tweaked. LinkedIn gets 40% because it got 40% last year and it worked, more or less Google gets 20% because… it’s Google etc. It looks and behaves like a plan. But it doesn’t bring the results that a solid plan would.
And look, last year's split is not worthless. If you've been running paid for a while, your own account numbers are the most useful thing you own. The trouble is treating that split as the ceiling.
But would you be able to estimate what 25% of the budget on Linkedin would be able to generate? What return are you even expecting from that Google spend? No? Then that sounds like a very expensive guess, considering you’re allocating around six or seven figures.
You also get to see how your numbers sit against benchmarks for companies and buyers like yours. A $340 cost per SQL feels fine in isolation. Against the benchmark for your segment, it's either excellent or it's the reason your pipeline is short, and you currently have no way of knowing which.
Count how many times you've rebuilt the budget spreadsheet from scratch. Not just updated it, but completely rebuilt from a near-blank tab every time a number changed.
New quarter, or a surprise cut in May, and off you go again: rebuilding tabs and re-linking cells, then trying to remember why last time's version assumed a 3% conversion rate here and 1.5% there. Half the logic lived in your head and left when you closed the file. So you reconstruct it, slightly differently, and quietly hope this version holds?
This costs you a day or three every cycle and pays you nothing back, because next time you start from… roughly zero again. A good plan should get smarter every quarter, with last cycle's actual numbers feeding next cycle's assumptions, so the model tightens over time. A spreadsheet you rebuild from scratch can't do that.
A proper planner should feel like an asset instead of a chore. So what does planning look like when it isn't a spreadsheet held together with hope?
There's a comforting myth in paid media: you can't really predict it, you just have to launch it and adjust as you go. Testing is real and it matters. But "we'll find out once it's live" is an expensive way to discover a channel was never going to work for your ICP, wouldn’t you say?
Good news is that with Harry, you can model a lot of this before a penny moves. Feed in the budget, and a decent planner returns a channel mix with projected results for each one, per budget level, based on Envy’s benchmarks for your kind of company and buyer. Call it a defensible starting position rather than a crystal ball, so you spend into a plan instead of a hunch 😉
The magic ingredient is… Envy’s benchmarks! A projection is only as good as the numbers behind it. Envy’s benchmarks are drawn from real campaigns for companies and audiences like yours, so the model reflects your world.
With that in place, the questions like, Why is LinkedIn getting this much? Because at this budget, the model projects a better cost per qualified lead there than the alternatives. Move budget between channels and you watch the projected outcome shift, before you commit, not three months and a blown quarter later.
A paid media plan is meant to be exactly that: a model that tells you the likely return on each pound before you spend it. Which leaves the fun part, actually deciding the split.
A plan is only as good as the parts you feed it. Three things to get right before you spend a penny:
Getting the targeting right is half the job. Your exclusions are the other half, and where most budget leaks:
Google spend lives or dies on structure and on what you exclude:
Each stage needs somewhere to send the clicks, or the budget has nowhere to go:
Don't hold the funnel too sacred though. People rarely move through it in a tidy line, they'll catch a blog post, then a competitor ad, then your demo page, in whatever order real life hands them, and you can't control the sequence. It's also why you shouldn't read too much into attribution metrics: the touch that gets the credit is never the only touch that did the work, especially when you consider the scary stats that tell you people need on average 76 touch points before they’re ready to speak with you.
Audiences, exclusions, keyword buckets, negatives, and a budget split across the funnel: a lot of moving parts to get right by hand, every quarter. Harry does them in one place.
Harry turns your annual or quarterly budget into a channel mix and projected results you can actually defend, before a penny goes out the door. Built on real benchmarks and custom to your business, and it's yours to keep. Next quarter or next fundraise, you just rerun it. Fixed scope and a clear end date, no retainer.
Prefer the plan built and run for you, not just handed over? Ask us about William. William is the full-service version, where our strategists take the plan live and manage the campaigns, then hand it back when it's working.
Either way, you stop rebuilding that spreadsheet at midnight. Give your budget a model it can stand on.