i run paid media as a pipeline system, not a media buy.

most B2B paid media reporting answers the wrong question well. it reports cost per lead to two decimals. it cannot say whether those leads became revenue.

the reason is structural. the ad platform can only see what it is told, and what it is told is usually a form fill. a form fill is not pipeline. in a long B2B cycle, the gap between those two is where the budget decision lives.

so the first measurement question is not which campaign is cheapest. it is which campaign produced contacts that reached a stage a salesperson cared about.

that reframing changes the buy. campaigns that look expensive on cost per lead often look correct on cost per SQL. the cheap ones frequently turn out to be buying job seekers and students.

i have seen exactly this on a live account. Kahuna Workforce Solutions had active Google Ads spend and no view of pipeline. broken URL tracking was silently breaking attribution in HubSpot. the team could not tell which campaigns produced demos, or whether spend worked at all.

once tracking was correct, the account was legible. $100,789 in year-to-date Google Ads spend. 345 tracked contacts and 54 demo requests. 30 reached MQL or beyond, 13 reached Sales Accepted, 7 reached SQL. verified as of July 21, 2026. that is a funnel you can make a budget decision against. cost per lead alone is not.

nothing in a paid account can be optimized past the quality of its conversion definitions. that is the whole reason this section comes before the channel sections.

the first job is deciding what a conversion is. in B2B that is rarely the thing the account is currently optimizing toward. a whitepaper download and a demo request are not the same event. a bidding algorithm told to treat them as equal buys the cheaper one forever.

so: one primary conversion per campaign. pick the closest signal to revenue with enough volume to train on. everything else is a secondary conversion, tracked but not bid on.

the second job is CRM truth. the CRM contact record is the ledger. the ad platform is a claim about that ledger. when they disagree, the CRM wins, and the disagreement itself is the finding.

the third job is reconciliation. GA4, the ad platform, and the CRM will each report a different number for the same event. that is normal. what is not normal is nobody knowing why. that is how a tracking break survives months.

the Kahuna case is the clean example. the spend was real and the campaigns were running. the failure was upstream of any optimization decision. URL tracking was broken, so HubSpot could not attribute. no campaign-level judgment was possible.

more on why this layer comes first. see [attribution](/insights/b2b-marketing-attribution).

or see [the mql problem](/insights/why-we-dont-measure-mqls).

the two channels do different jobs and should not be judged on the same metric.

google ads b2b saas is intent capture. someone is typing a problem, and the account exists to be there with the right page. the structure question is mostly negative keywords and match type discipline. broad match in a niche B2B category burns the budget on adjacent industries in a week.

the campaign layout follows intent tiers. high-intent product and competitor terms get their own campaigns and budget floor. problem-aware terms get a separate campaign with a lower bid and a different landing page. research terms get very little, or nothing.

linkedin ads b2b is demand creation against a defined audience. nobody on LinkedIn is searching for you. you are choosing a list and paying to be in front of it repeatedly.

that makes the LinkedIn measurement window longer and the conversion event softer. holding LinkedIn to branded search cost per SQL is a category error. it kills the channel before it can work.

budget by segment, not by channel. decide what the ICP tiers are, then allocate to the tier and let the channel serve it. an account structured by channel budget will always underfund the segment that needs both.

creative and landing pages are part of the media buy, not a separate project. the most common failure I see is a good account pointed at a generic homepage.

the two terms get used interchangeably. they describe different systems with different reporting.

lead generation optimizes for the volume of contacts entering the database. it is measurable quickly and it is easy to scale. its failure mode: volume and quality diverge once you push past a natural ceiling.

demand generation optimizes for the number of qualified accounts that arrive ready to buy. it is slower to measure and it includes work that produces no immediate form fill.

the honest version is that most B2B SaaS companies need both, sequenced. lead generation captures the demand that already exists. demand generation creates the demand that does not.

the reporting difference matters more than the definitional one. a lead generation program judged on pipeline will look like it failed. a demand generation program judged on cost per lead will look like it failed. the metric has to match the job.

this is also where the MQL problem lives. an MQL is a scoring convention, not a fact about a buyer. treating it as a shared target between marketing and sales is what starts the blame loop.

the b2b paid media agency model is built around media management. it does that part well. the boundary is where the contract ends.

an agency sells account management. campaign build, bidding, creative, and platform reporting. that is real work and a competent agency will beat an untrained in-house hire at it.

the structural limit is simple. conversion definitions, the object model, and attribution sit inside the client.

an agency can recommend a change to any of the three. it usually cannot make one.

that gap is exactly where the Kahuna failure lived. the media management was not the problem. the tracking upstream was. it stayed broken because it sat on the client side of the line.

an operator installs across that line. same campaign work. plus conversion definitions and CRM reconciliation. plus the authority to change the primary conversion when it is wrong.

the test is the same one question as everywhere else. when paid pipeline drops next month, who diagnoses it. and do they have CRM access.

the first thirty days are a measurement rebuild and the first spend corrections. no new campaigns in week one.

week one is evidence. i connect the ad accounts, GA4, and the CRM. then reconcile on one event and date range. the gaps between the three are the finding.

week two is conversion definitions. what is the primary conversion, what is bid on, and whether those are the same event. this is usually where the largest single correction is.

week three is waste and structure. search terms, negative keywords, audience overlap, budget concentration, message match.

week four is the corrected buy. budget reallocated by segment and primary conversions set. the reporting line is built so next month is judged on stage progression.

want the diagnosis without the install. that is the [growth diagnostic](/growth-diagnostic). same first two weeks, delivered as a ranked list with a dollar figure per item.

the accounts are yours and stay yours. I work inside your ad accounts, analytics, and CRM. never a reseller account that holds your history hostage.

you own the conversion definitions, attribution method, and campaign structure. also the negative keyword lists and the reporting method. written down, not held in my head.

the part worth the most is the reconciliation procedure. tracking breaks, and it breaks silently. a documented monthly reconciliation turns a six-month silent failure into a one-month one.

if the engagement ends, the account keeps running. the reporting produces the same number. that is the bar.

paid media is priced as a fixed-scope engagement, never as a percentage of ad spend.

percentage of spend is a direct conflict of interest. it pays more for spending more. the correct call is often to cut a campaign to zero.

current scopes and figures are on [pricing](/pricing).

start with the diagnosis. thirty minutes, live, on your own ad accounts and CRM, not a template audit.

you leave knowing what is broken between the click and the CRM record. and what it costs. if there is a fit, I scope the install on the spot. if there is not, the list is still yours.

book it from [growth diagnostic](/growth-diagnostic).