A B2B SaaS client in the HR tech space came to us spending $28,000 per month across Google Ads and LinkedIn, with a blended cost per lead of $412. Their in-house team was optimising toward last-click conversions, which meant LinkedIn was getting almost no credit despite being the channel where 60% of closed deals had their first touchpoint. After rebuilding their attribution model and reallocating budget based on actual influence rather than final click, their CPL dropped to $189 within three months.
The Starting Problem: Last-Click Was Lying to Them
When we audited the account, Google Search was receiving credit for nearly every conversion because it was the last click before a demo request. LinkedIn campaigns looked expensive at a surface-level CPL of $680 and were being gradually defunded. But when we pulled CRM data and matched it against ad exposure logs, LinkedIn-assisted deals had a 34% higher close rate and a 22% higher average contract value than leads attributed to Google alone.
Last-click attribution is structurally biased toward bottom-funnel channels. A prospect who sees three LinkedIn ads over six weeks, reads two blog posts via organic search, and then converts on a branded Google query is recorded as a Google conversion. The LinkedIn spend looks wasteful, gets cut, and pipeline volume drops a quarter later. This is a pattern we see consistently, and it is documented in detail in our article on multi-touch attribution for B2B ROI.
The fix starts with connecting your CRM to your ad platforms using UTM parameters that survive across sessions, not just the last session. For this client, that meant implementing a first-party data layer via their HubSpot CRM and mapping each contact's full touchpoint history before any budget decision was made.
Attribution Model: What We Built and Why
We moved the client from last-click to a position-based (U-shaped) model, giving 40% credit to the first touch, 40% to the lead-conversion touch, and distributing the remaining 20% across assisted middle touches. This is not the most sophisticated model available, but for a sales cycle of 45-90 days with two to five touchpoints per deal, it reflects reality far better than last-click and is easy enough for their marketing director to explain to the CFO.
We used Google's data-driven attribution documentation as a reference for setting up conversion actions correctly in Google Ads, then cross-referenced those signals with HubSpot's attribution reports. The key technical requirement was ensuring every LinkedIn and Google ad URL carried consistent UTM parameters that matched the naming convention inside HubSpot, which sounds obvious but was completely broken in their existing setup.
We also set up a separate 'influenced revenue' column in their reporting dashboard, pulling closed-won deal values back into a spreadsheet segmented by first-touch channel. This gave the client a single source of truth that both the paid media team and the sales team could trust.
Budget Reallocation: Where the CPL Drop Actually Came From
Once attribution was working correctly, the numbers told a clear story. LinkedIn was generating first-touch influence on 58% of closed deals but receiving only 29% of the total paid budget. Google Branded was receiving 18% of budget and generating zero first-touch influence, because by definition, if someone is searching your brand name they already know you. We cut branded Google spend by 60% and shifted that $5,040 per month into LinkedIn Conversation Ads targeting a custom audience of the client's ICP job titles at companies in their target revenue band.
The Google Non-Brand campaigns were restructured to focus purely on high-intent, bottom-funnel keywords, with tighter match types and a refined negative keyword list. If you are running broad or phrase match on competitive SaaS keywords without an aggressive negative keyword strategy, you are paying for clicks that will never convert. Our guide on eliminating wasted spend with negative keywords covers the mechanics of that cleanup in detail.
Within 60 days of the reallocation, LinkedIn-sourced demo requests increased by 41%. Google CPL dropped from $290 to $198 because the remaining budget was concentrated on higher-intent queries. Blended CPL across both channels fell to $231 at the 60-day mark and continued down to $189 by month three as the LinkedIn campaigns accumulated data and the algorithm optimised toward the client's conversion audience.
What the Lead Quality Data Showed
Lower CPL only matters if lead quality holds or improves. In this case, quality went up measurably. The SQL-to-demo ratio improved from 38% to 51%, meaning more of the leads coming through were genuinely qualified. The sales team reported that inbound prospects were arriving with more context about the product, which is consistent with LinkedIn's role as a mid-funnel education channel: prospects had already seen multiple pieces of content before requesting a demo.
Pipeline velocity also shortened. Average time from lead to closed-won deal dropped from 74 days to 61 days over the three-month period. That improvement is hard to attribute entirely to the attribution and budget changes, since the client also ran a product launch during this window, but the directional correlation is strong. When your first-touch channel is doing real qualification work, the sales cycle gets shorter because you are not spending the first two calls explaining basic product-market fit.
According to Gartner's research on the B2B buying journey, buyers complete a significant portion of their evaluation before engaging a vendor. Investing in channels that intercept buyers early in that process, and then measuring their contribution properly, is where the real efficiency gains in B2B paid media come from.
Three Things to Replicate This Result
This outcome was not the product of a single clever tactic. It came from fixing three compounding problems at once: broken attribution, misallocated budget, and disconnected CRM data. If you are running paid search and paid social together without connecting your ad data to your CRM at the deal level, you are almost certainly making budget decisions based on incomplete information. The channels that look expensive in Google Analytics or in-platform reporting are often the channels doing the heaviest lifting in your actual pipeline.
- Audit your UTM parameter consistency across every paid channel before drawing any attribution conclusions.
- Pull closed-won deal data from your CRM and segment by first-touch channel, not just last-touch. The gap between the two numbers is your misallocation estimate.
- Run a 30-day parallel period where you report both last-click and position-based attribution side by side before changing any budget. This builds internal confidence in the new model.
- Set a minimum budget threshold on any channel that influences more than 20% of your pipeline, regardless of what in-platform CPL looks like.
- Review negative keyword coverage on all Google Non-Brand campaigns before scaling. Wasted spend on irrelevant clicks inflates CPL and poisons Quality Score data.
If you are working through a similar audit and want to understand how your Google Ads structure might be contributing to poor lead quality independent of attribution, the breakdown in our article on why your B2B landing page is not converting covers the post-click side of the same problem. Attribution fixes the measurement layer, but the conversion layer has to work before any of it matters.