Someone on your team asks it in every Monday meeting: what is actually working in marketing?
You open your analytics. Search sent the most visits. A newsletter spiked last Thursday. Social got plenty of likes.
But nobody can say which of those turned into a paying customer, so the meeting ends with opinions instead of answers.
This guide shows you why it is harder in 2026 than it was a few years ago, and how to do it without tracking cookies.
It is written for founders and small marketing teams who want a clear answer, not a 40-tab attribution model.
Why is “what’s working?” so hard to answer?
Your customers meet you in more places than your analytics can follow, and the tools that used to fill the gaps have changed.
Five things get in the way:
- Budgets are tight, and results are under review: Gartner’s 2026 CMO Spend research found marketing budgets on a plateau since 2022, at 7.8% of company revenue. Fewer CMOs report beating their targets, and more are missing them on acquisition, retention, and ROI. One of Gartner’s recommendations is to measure full-funnel impact beyond last-click attribution.
- Clicks are disappearing: SparkToro and Similarweb found that 68.01% of US Google searches ended without a click in the first four months of 2026, up from 60.45% in 2024. A visit is a weaker sign of influence than it used to be.
- Browsers limit what scripts can remember: Safari deletes cookies created in JavaScript after 7 days without user interaction with the site, and caps them at 24 hours when it detects link decoration used for cross-site tracking, according to WebKit’s tracking prevention documentation. Google kept third-party cookie choice in Chrome and retired its Attribution Reporting API in October 2025, so measurement still has to come from your own first-party data.
- Buyers decide before they talk to you: In 6sense’s 2025 Buyer Experience Report, first contact with a vendor happened about 61% of the way through the buying journey, down from 69% a year earlier. Much of the influence happens where no click is recorded.
- Default reports favor the last touch: GA4 offers three attribution models, and all of them exclude direct visits from receiving credit unless the whole path was direct. First click, linear, time decay, and position-based were removed in November 2023.
None of this means you cannot measure. It means you should measure the part you can trace, and ask about the rest.
What does “proving marketing ROI” mean for a SaaS?
It means showing, for each channel, how many paying customers it produced and what they cost to win.
For an early or mid-stage SaaS, that takes three numbers per channel:
- Paying customers: Not signups, not trials.
- Visit-to-paid rate: Paying customers divided by visitors from that channel.
- Cost per paying customer: Ad spend plus the tools, freelancers, and hours you put into the channel.
The formula is simple:
Channel ROI = (revenue from that channel’s customers - channel cost) / channel cost
Use first-year gross profit if you have it. Use monthly recurring revenue if you do not.
You do not need multi-touch modeling to answer “which marketing channel converts?”
You need a reliable first source and a revenue event you trust. Everything else is refinement.
Can you prove marketing ROI without cookies?
Yes, with one trade-off. Most of what you need does not depend on cookies at all.
What works without cookies:
- Source and UTM data for each session
- A user ID at signup that merges the anonymous session
- Revenue events sent from your server
- The self-reported answer at signup
What gets harder: following an anonymous visitor across several days before they sign up.
If you need longer memory, Vemetric supports cookies as an opt-in.
Turn them on only after you have collected consent from your visitors, and check the rules that apply to you with your legal counsel.
To compare approaches, read our guides to cookieless tracking solutions and GDPR-compliant product analytics.
What does Vemetric cover, and what does it not?
Vemetric is web and product analytics for SaaS teams. It is not a full attribution suite, and it helps to know the difference before you start.
What you get:
- Referrers, referrer types, and UTM tags on a real-time dashboard
- User journeys with a first-seen page and referrer
- Anonymous-to-identified merging at signup
- Funnels and custom events, with backend SDKs and a REST API
- Cookieless tracking by default, EU hosting, and an open-source codebase under AGPLv3
What it does not do today: multi-touch credit modeling, ad spend import, or CRM sync. Campaign insights are on the roadmap.
For most SaaS teams with a handful of paid channels, first-touch data plus revenue events plus the self-reported answer is enough to make good decisions. When you outgrow it, our roundup of marketing attribution tools covers the next step.
What should your monthly ROI report include?
Keep it to one page your team can read in two minutes:
- Paying customers by first source
- Visit-to-paid rate by channel
- Cost per paying customer by channel
- The top five self-reported answers
- Change from last month
- One decision you will make because of it
Then you can answer the Monday question in plain words:
“Last month, SEO content brought 21 of our 37 new customers at about $29 each. Paid ads brought 6 at $250. Self-reported answers point to podcasts, and Slack mentions we cannot see in clicks. We are moving $500 from paid to content.”
FAQs
Neither tells the whole story. For a small SaaS, first-touch data plus a self-reported answer is usually more useful than last-click, because it shows what created demand. GA4 now offers only data-driven and last-click models, and none of them credit direct visits.
Not at first. Multi-touch tools help when you run several paid channels with long sales cycles and enough data to model. Until then, a clean first source and a trusted revenue event answer most “what’s working?” questions. Compare options in our marketing attribution tools guide.
