John Wanamaker's line about wasted ad spend is over a hundred years old. Marketing analytics has come a long way since then. And yet: only 54% of marketers are confident measuring ROI across digital channels. That should sting.
The problem isn't data. Most marketing teams have too much of it. The problem is that most measurement setups aren't designed to answer the question that actually matters: is this marketing activity producing qualified pipeline?
Here's what needs to change, and where to start.
The Real Reason Measurement Fails
Most measurement problems start before anyone opens a dashboard. They start when the goal is set.
"Increase brand awareness" is not a goal. Neither is "generate more leads." They're directions, not destinations. Without a specific number and a deadline attached, there's no way to know whether you're winning or just moving.
This isn't pedantic. Research from CoSchedule found that marketers who set specific goals are 377% more likely to report success than those who don't. That's not a marginal edge. It's a different game.
A useful goal looks like: "Generate 40 MQLs per month from paid channels by Q3." You know the metric, the volume, and the timeline. Every decision on budget, creative, and channel can be tested against that target.
Make Your Goals Connect to the Business
Marketing goals that float in their own universe don't serve anyone.
If the business needs to close 9 new clients this year, your goal structure should trace directly back to that number. Work backwards. Nine new clients. At what close rate from sales-qualified leads? That gives you an SQL target. At what MQL-to-SQL conversion rate? That gives you an MQL volume. At what cost per MQL from each channel? That gives you budget allocation.
It's not complicated. Most teams just don't do it because it requires knowing the sales numbers, which requires actually talking to sales. The measurement conversation is often where the sales and marketing alignment conversation should have already happened.
Choosing the Right KPIs
Once you have goals, KPIs follow. The trap is tracking too many of them.
Pick 1-3 KPIs per goal. If your goal is pipeline generation, your KPIs are MQL volume, cost per MQL, and MQL-to-SQL rate. Not impressions. Not follower count. Not email open rates unless those emails are directly tied to pipeline movement.
The vanity metric problem is real. Page views and social likes feel like progress. They're often not. A piece of content can pull 10,000 visits and generate zero qualified leads. Until you track what happens after the click, you're measuring the wrong thing.
That said, not every metric needs to tie directly to revenue. Leading indicators matter.
Organic traffic, for example, doesn't convert today. It converts in 60 to 90 days. If you only track short-term conversions, you'll underfund the channels that fill the top of the funnel and wonder why pipeline dries up six months from now.
The best measurement setups carry both: short-term conversion metrics and longer-range indicators that predict where pipeline will come from next quarter. This balance between SEO and demand generation is where most B2B teams leave performance on the table.
Where Attribution Actually Breaks
Here's where most B2B measurement goes wrong: last-click attribution.
Last-click gives 100% of the credit for a conversion to whatever the buyer touched right before they filled out a form. Usually, that's a branded search or a retargeting ad. So the model concludes: paid search is working, retargeting is working, that blog post and LinkedIn campaign from six months ago did nothing.
That conclusion is probably wrong.
In B2B, buying cycles are long. A VP of Marketing at a manufacturing company might read your blog in January, attend a webinar in March, download a case study in April, and finally request a demo in May. If you're only measuring last-click, you've credited the demo confirmation email and written off the other four touchpoints.
Multi-touch attribution models distribute credit across the full journey. Linear models split it evenly. Time-decay models weight recent interactions more heavily. Position-based models give the most credit to the first and last interaction, with middle touches getting a smaller share.
No model is perfect. But any multi-touch model gives you a more accurate picture than last-click alone.
If you want harder evidence, run hold-out tests. Turn off a channel in one region or segment. If performance doesn't change, the channel wasn't doing what you thought. eBay did exactly this with paid search and found it was largely cannibalizing organic traffic. They were paying for clicks they'd have gotten anyway.
Using the Data to Actually Make Decisions
This is the part most teams skip. They build dashboards, share numbers in meetings, and then nothing much changes.
Data is only useful when it changes what you do next.
If your cost per lead on one channel is three times higher than another, shift budget. If a content piece drives traffic but no conversions, look at the landing page, the CTA, the offer. If MQL volume is up but SQL rate is dropping, you have a lead quality problem, which usually means your targeting is too broad.
You don't need perfect data to act. Good marketing teams move on directional data and tighten the model as they go. The teams that wait for certainty before optimizing are always behind. As MarketingProfs notes, velocity of measurement and decision-making matters more than waiting for a perfect data set.
A few things worth building into your regular cadence:
- Weekly: Check KPIs against targets. Flag anything moving in the wrong direction early.
- Monthly: Review attribution data and channel mix. Look for trends and anomalies.
- Quarterly: Revisit goals against business outcomes. Reforecast if the numbers warrant it.
Common Mistakes Worth Skipping
Tracking more than three KPIs per goal. The more you track, the less you focus. Pick what matters most.
Comparing periods without accounting for seasonality. A 20% drop in MQLs in August means almost nothing without knowing what August looked like last year.
Treating attribution as settled. No model tells the full story. Use attribution as evidence, not verdict. Combine it with experiments when the numbers are telling you something surprising.
Conflating correlation with causation. Sales went up the same month you ran a campaign. Maybe the campaign drove it. Maybe there were three other factors. The only way to know is to test.
Getting the Setup Right
The tools matter less than the process.
Google Analytics, HubSpot, your CRM, and a clean UTM tagging system will cover most of what you need. What matters more is that the tools are set up correctly, the data is accurate, and someone is actually looking at it on a regular schedule.
If conversion tracking is broken or lead source data is unreliable, no dashboard is going to fix that. Start with data quality. Everything else builds from there.
The Bottom Line
Measurement is only useful if it's connected to what the business is actually trying to do. Set goals that tie to pipeline. Pick KPIs that reflect those goals, not the ones that are easiest to report. Build attribution that accounts for the full buying journey. And act on what the data is showing you, even when it's inconvenient.
The goal isn't a better dashboard. It's a clearer view of what's working, so you can put more budget there and stop wasting it where it isn't.
If you're not confident your current measurement setup is giving you an accurate picture of what's driving pipeline, that's worth resolving before your next budget cycle. Talk to the NP team about building a measurement system that connects your marketing activity to actual sales outcomes.
Frequently Asked Questions
Resources
