Ask a room of marketers which channel is working and you'll get a fast answer. Ask why, and it gets quieter. Paid media tends to get handled like a settings problem, the kind of thing a sharp media buyer fixes from inside the platform, and every so often that's exactly what it is. Usually it isn't. Usually the campaign is doing precisely what it was built to do, and nobody went back to check whether that thing was ever wired to revenue.
That's roughly the territory our founder Matthew Simpson spent an hour in on episode 66 of The Marketing Share, the podcast Alec Cheung and Barb VanSomeren built for business leaders who inherited a marketing function they didn't come up through. Matthew's worked in-house and on the agency side, and the two sides have a habit of describing the same failure in completely different vocabularies.
Agencies and marketing leaders aren't solving the same problem
An agency walks in with channel expertise and a media plan. The person across the table has a number to hit, a sales team that's stopped trusting the leads, and a budget that has to survive the next quarterly review. Both sides think the meeting is about paid media. It isn't, quite, and a plan that nails the first job while ignoring the second one tends to get killed around month six whatever the CPL is doing.
Paid works in B2B, just not on ecommerce math
The skepticism is earned. Plenty of B2B teams have put real money into paid search and paid social, watched it come back as form fills that sales quietly ignored, and concluded that paid doesn't work for them. They're reporting the experience accurately. It's the conclusion that goes sideways.
What changes in B2B is the shape of the return rather than whether there is one. You're selling to a committee, the cycle runs long, and a lot of the influence never leaves a click behind, so a channel can be pulling real weight and still look mediocre in a last-touch report. Judge a nine-month sales cycle on a 30-day conversion window and you'll switch off the thing that was building demand while you keep funding the thing that was harvesting it.
When it breaks, start a floor or two above the account
This is the part that matters if you're the one holding the budget. Performance drops, and the instinct is to open the account and start turning things, because those levers are visible and turning them feels like progress. The cause is usually further up. Maybe the offer doesn't land with the segment you're targeting. The landing page might be asking for a commitment nobody makes on a first visit. Or marketing and sales quietly stopped agreeing on what counts as a qualified lead, and nobody wrote down the week it happened.
Bid adjustments don't touch any of that. Matthew has put it about as plainly as it can be put on our own site: “Most marketing fails not because of bad ideas, but because of bad structure. We build the structure first, strategy, channels, measurement, and then we move fast.”
Automation isn't the same thing as leaving it alone
Every platform sells automation as permission to stop thinking about the account, and every account that stops getting thought about starts to slide. Audiences saturate, competitors move their bids, creative wears out, and the algorithm keeps optimizing toward whatever signal you handed it long after that signal stopped predicting revenue.
Smart bidding is good technology and it's also completely indifferent, so it'll chase a bad target with the same efficiency it chases a good one. The teams who get durable results out of paid run it as an operating cadence instead of a launch, which mostly means somebody is in there on a rhythm asking what moved and why, then feeding better signal back in.
Attribution won't settle the argument for you
Multi-channel attribution promises a clean answer about which channel earned the credit, and it can't produce one, because the data underneath has holes no model fills. Deals move on dark social and on a recommendation in somebody's Slack group, neither of which leaves a trace, and cookie loss and privacy changes have made even the trackable paths shakier than the dashboard admits.
You still need to measure. You just shouldn't let the model hand down verdicts. Self-reported attribution on the form, holdout tests, and watching whether pipeline actually moves when spend moves will get you nearer to cause and effect than a weighted model will, at a fraction of what the model costs to build and keep alive.
A small budget forces a sequencing decision
Small budgets often get run like scaled-down big ones, spread thin across the same channel mix so nothing gets neglected. What that produces is a portfolio where no single channel ever sees enough volume to throw off a readable signal, and twelve months later you still can't say which part worked.
Concentration usually beats coverage. Pick the channel where your audience is easiest to reach and your offer is most believable, fund it past the point where the numbers start meaning something, learn from it, and add the next one after that. It feels riskier than spreading out, and it's the only version that teaches you anything inside a fiscal year.
This ladders up further than the marketing budget
The back end of the conversation makes a case that's easy to scroll past, which is that performance marketing belongs in the value column rather than the cost column. A company that can acquire customers repeatably at a known cost is worth more than a company with identical revenue and no account of where it came from. Boards and buyers pay for predictability, and a marketing function that can show the mechanism behind the number rather than just the number is an asset on the balance sheet.
Where that leaves you
If paid isn't performing, hold off on the settings for a minute. The more useful question is whether the strategy, the offer, the audience definition and the measurement sitting underneath that campaign are solid enough to make optimizing it worth the hours. Get that right and the channel work gets a lot easier. Skip it and you can run a technically flawless campaign against a target that was never going to move the business.
Watch and listen
Watch the full conversation on YouTube, or listen to episode 66 on The Marketing Share.
If you want a read on where your own paid program is leaking, that's what our performance marketing work is built for, and you can book an intro call whenever you're ready.