Paid media management
Google Ads, Microsoft Advertising, Meta and LinkedIn run under one retainer for manufacturers and other B2B companies selling considered purchases, the kind where a buyer researches, asks for a quote, talks to a person, and decides months later. I build the campaigns and I run them. The reporting starts with what your CRM says closed.
The problem
Most accounts I take over are already reporting well. The platform shows a healthy cost per conversion, the monthly deck is full of green arrows, and nobody in the building can tell you how many of those conversions became customers.
Google counts a phone call, a form submission, and a duplicate of that same form as three conversions. Auto-tagging takes credit for people who already knew your name and typed it into the search bar. Meanwhile your CRM, the system where your sales team tracks leads and deals, files call-tracking leads with no source attached, so real calls driven by paid search look like they came from nowhere.
Both scoreboards are wrong, in opposite directions. Until somebody reconciles them, every bid, budget and pause decision in the account is being made on numbers nobody has checked.
The gap nobody reconciles.
The work
One retainer, all four platforms, no separate line items for the parts that make them work.
Accounts get segmented by the kind of buyer you are selling to rather than by engine or by branded versus non-branded. Those internal splits still matter to me, they just are not how your team makes decisions.
Target return on ad spend is set from your gross profit per sale, so the bidding math breaks even where your business breaks even. Conversion actions get valued from observed CRM close rates and average deal size, because a form lead and a phone call almost never convert at the same rate and the platform will happily treat them as equals.
Regular review of which queries actually spent money, with every candidate negative keyword checked against the CRM before it gets cut. A term that looks like pure waste in the platform sometimes has closed deals sitting behind it.
Written for someone working through a specification or a purchase committee, not for an impulse buy. Tested where the result can be judged on leads and sales rather than on a platform quality score.
Monthly budgets built to real search demand for your core terms and staggered across the season, instead of a flat twelfth of the annual number every month. Spend lands when your buyers are in market and backs off when they are not.
Inside the retainer
Media buying is the visible part of the job. These three sit inside the same retainer, because handing them to somebody else is how accounts stall for a quarter while two vendors email each other.
Nothing gets optimized until I know the conversions are real events and the CRM can see where a lead came from. On an industrial manufacturer's account, the CRM was filing phone-call leads with no source attached, so paid search was being judged on a lead count that was missing its own calls. Matching ad-platform call records to CRM contacts by area code and timestamp identified them, and the same match was run backwards over the archive to reclaim the ones already lost.
When the page receiving the traffic is the constraint, I rebuild the page rather than optimize around it. That covers the copy, the offer, how much the form asks for, and the proof a considered-purchase buyer needs before they will give you a phone number.
Paid and organic compete for the same queries and land on the same pages, so I keep them working from one plan. On a leadership development firm, fixing the FAQ content and page metadata moved one landing page and its most important query.
The call-tracking fix above is a reporting overlay built from per-call data, not a bulk rewrite of anyone’s CRM. The durable version, a dedicated tracking number per channel that writes its own source, went in afterwards. Paid search had been running about 14% more real lead volume than anyone was crediting it with.
SEO here is a supporting capability, not a second retainer. If what you need is a full standalone SEO program, I will tell you that on the call and point you at someone who does it properly. If your question is whether AI assistants recommend you when a buyer asks them for options, that is separate work and it has its own page: AI Visibility Audit.
Measurement
On an industrial manufacturer, I compared what the ad platforms reported against what actually reached the CRM over the same three months. One platform's search campaigns reported 63% more conversions than there were real contacts. The other overstated by 44%. Same account, same period, both wrong in the same direction.
Nobody had caught it, because nothing about it looks broken. The number goes up when you spend more, which is exactly what a healthy account is supposed to do. It only surfaces when somebody counts the leads in the system where the business actually keeps them, and that is the number I report on. Drift like this runs in both directions depending on the campaign, so there is no correction factor to apply. You have to go and check.
Conversions reported against leads in the CRM
One industrial manufacturer's search campaigns on a single platform, three consecutive months in 2026, indexed with the reported figure set to 100. A side-by-side comparison of two counts of the same period, not a change over time.
How results get counted
Platform conversions are an input, not a result. The scoreboard is what your system records: leads a salesperson actually touched, quotes that went out, deals that closed, revenue collected. When I report cost per lead or cost per customer, it is spend divided by outcomes in your CRM, never spend divided by conversions counted in the ad platform.
I work in HubSpot daily. Beyond that the brand does not matter much: if your system can export a lead with a source and a date, or hand me read-only access, the method is the same. That includes the spreadsheet a sales manager actually maintains, which is more common in manufacturing than software vendors like to admit, and it works fine as a starting point.
Gross profit is the other half of it. The same return on ad spend can be a comfortable win for one business and a slow loss for another, depending entirely on margin. So targets come off your gross profit per sale and your close rate, which is also how we find out early whether a channel can pay for itself at all. If the math says it cannot, that is a finding worth having in week one rather than in month six.
And when the data is genuinely ambiguous, the report says so in plain words. A smaller number I can defend to the decimal is worth more to you than a big one that comes apart the first time your CFO pushes on it.
An urgent care group with several locations was reporting about 4,000 conversions a month. Most of those were page views and clicks on driving directions, counted as if somebody had booked. I went to the event level, rebuilt the count around the actual booking confirmation, and then verified the rebuild: over 60 days, 5,505 events, 100% of them firing on real booking confirmation pages across every one of its locations.
One honest note on that comparison. Our figure counts completed online bookings tracked in Google Ads. The corporate benchmark counts confirmed visits in their medical records system. The two are not counting the same event, so read the comparison as directional rather than exact.
Results
One account, five years in. Most recent full year against the prior full year, attributed in the client’s own CRM rather than in the ad platforms.
The thesis that year was to segment harder, stop paying for the traffic that had never once closed, and give that budget to the segments that do. We did not set out to spend less. It came out that way.
Who does the work
I am Andre Rosdahl. I build the campaigns, I do the analysis, and I am the person on the call when a month goes badly.
My longest relationships run five and six years, with others past the two-year mark. Nobody stays that long for a monthly report.
Fit
Full detail on that, including what I need from you and what I turn down, is on how we work.
Questions
Month to month, cancel with 30 days’ notice. Considered-purchase B2B accounts take a real quarter to show their shape in the data, so I would not recommend judging the work at 60 days, but nothing here locks you in.
There is no published rate because it depends on your spend level and how much needs rebuilding on day one. You will get a real number on the intro call, not a follow-up proposal three weeks later.
Tracking gets verified and the account gets its first real changes inside the first 30 days (see how we work for the exact sequence). Meaningful movement in cost per customer usually shows up over a full quarter, because considered-purchase sales cycles do not compress just because the ads improved.
Some clients keep an in-house marketer and bring me in for the parts that need a specialist: platform execution, tracking rebuilds, and the CRM reconciliation most in-house teams do not have time to build. I will tell you on the call if your team already has this covered.
Access to the ad accounts, read access to the CRM, and a working definition of what counts as a real lead in your business. Without the CRM connection, reporting reverts to platform numbers, which is the exact problem this page describes.
It shows up in the report with the reason attached, and I am on the call to explain it. No quiet rewording of the metrics to make the arrow point up.
Tell me what you sell and what you are spending. By the end of the call you will have the three changes I would make first, whether or not you hire me to make them.
Book an intro call