How to measure what’s working in your marketing without an attribution platform

You don't need an expensive attribution platform to prove demand gen is working. A hybrid model combining free analytics tools with one simple form field, "How did you hear about us?", tells you most of what an enterprise attribution platform would, at a fraction of the cost. For lean marketing teams with tight budgets and little time, it's the smarter option.

Measurement in B2B marketing has a long-standing problem. A landmark Refine Labs study tracking 620 conversions and $21.5 million in closed-won ARR found a 90% measurement gap between what software-based attribution reported and what customers actually said influenced their purchase. Podcasts drove 53% of revenue ($11.4M) according to self-reported data but software attribution credited them with precisely 0%. Meanwhile, software attributed 82% of conversions to Google Search and direct traffic, yet customers self-reported those channels only 12% of the time.

This guide is for marketing leads with no analyst and little time. It covers the smallest setup that gives you a reliable picture: three things to set up once, four things to watch, and one hour a month to review them.

Why "good enough" measurement beats paid solutions for growing businesses

In B2B attribution, even the expensive tools get it wrong most of the time. Dreamdata's own 2026 benchmarks show the average B2B customer journey involves 88 touchpoints across 4 channels, with 10 stakeholders, over 272 days. No attribution model, not even one costing £15,000/year, captures all of that.

The philosophy here follows the 80/20 principle. The 20% of effort that delivers 80% of attribution value consists of four things:

1. Adding a self-reported attribution field to your high-intent forms,
2. Maintaining UTM discipline across your links,
3. Using GA4 and your CRM's free tier for digital tracking, and
4. Reviewing both data sets monthly.

The remaining 80% of effort (multi-touch modelling, AI-driven analysis, cross-device tracking, automated audience syncing, incrementality testing) yields diminishing returns that lean teams cannot justify, in both time and expense.

What paid platforms actually do that you'll be replicating. Dreamdata and HockeyStack provide account-level journey stitching, automated revenue attribution, IP-to-company identification, AI-driven insights, and real-time audience syncing to ad platforms. Those are genuinely valuable features for companies with the budget and the resource to analyse the data (and with having used them before, I can say they are great tools). This framework replicates their core insights: understanding what creates demand (not just what captures it), connecting marketing activity to pipeline and revenue, and identifying which channels deserve more investment. You'll lose the automation and dashboards but, importantly, you’ll keep the strategic clarity. There is one more reason these platforms are hard to justify at growth stage: they are built for teams running complex, high-volume paid programmes across multiple channels. If your paid spend is £2-3k/month on one or two channels, the attribution granularity they provide will not change your decisions. This framework is small enough to see clearly without them.

How privacy changes make this framework more relevant, not less

Tracking keeps getting less complete. Safari and Firefox block third-party cookies, a large share of visitors decline cookie banners, and the modelling GA4 uses to fill that gap needs more traffic than many smaller sites have. AI assistants add a new gap: GA4 added an AI Assistant channel in May 2026, but visits that arrive without a referrer still show up as Direct.

In practice, self-reported attribution has zero privacy or tracking dependencies. It doesn’t require cookies, consent banners, tracking pixels, or data processing agreements. Asking "How did you hear about us?" on a form is inherently compliant and unaffected by any privacy regulation. As software-based attribution becomes less reliable, the simple act of asking your customers becomes relatively more valuable. This is exactly why this framework was built.

What to measure, and what to stop measuring

The metrics that matter for growth

1. Qualified enquiries: people who asked to talk to you and fit who you sell to. This tells you whether marketing is creating buyers right now.
2. Branded search: how many people search for your name. Use the branded queries filter in Google Search Console. It only appears on sites with enough search volume, so if yours doesn't have it, track impressions on a branded Google Ads campaign instead.
3. Self-reported attribution reveals what's actually creating demand versus what's capturing it.
4. Revenue from enquiries: the value of deals that started as an inbound enquiry, and how much of it closed.

What to stop measuring

Several common metrics can rise while revenue stays flat. Stop treating these as measures of success: MQL counts, email open rates (Apple's Mail Privacy Protection inflates them), follower counts, raw traffic, cost per lead on its own, click-through rates on awareness ads and gated content downloads. Look at qualified enquiries and the revenue they produce instead.

The setup: three things to do once, then one hour a month

Ask buyers how they found you

Add a required free-text question, "How did you first hear about us?", to every form where someone asks to talk to you: enquiry, quote, demo, contact and pricing forms. On first calls, whoever speaks to the buyer asks "What first put us on your radar?" and logs the answer in the CRM. The next section covers how to get this right.

Keep the tracking basics clean

You don't need an analytics project. You need four things in place:

• UTM tags on every paid and external link, using one shared naming list. Never tag links within your own site, because that overwrites the original source.
• Your CRM recording the original source of every contact. HubSpot does this by default. If you're on HubSpot's free tier, it currently allows 10 custom properties across the whole account, so spend them carefully.
• GA4's AI Assistant channel, which since May 2026 sorts visits from tools such as ChatGPT, Gemini and Claude into their own line with no setup. It only catches visits that arrive with a referrer, so some AI-driven visits will still show as Direct.
• GA4's attribution lookback window set to the longest option available (Admin → Attribution Settings), so longer buying cycles aren't cut short.

Review it once a month

Book one hour on the first Monday of each month. Update a simple sheet with four columns (month, branded search clicks, qualified enquiries and revenue from enquiries). Then read every "how did you hear about us" answer from the month and compare them with what GA4 reports.

When the software credits Google and the buyer names a podcast, a colleague or an AI assistant, both are right. The software shows where they clicked. The buyer tells you what made them look. Channels that buyers mention but the software barely credits are the ones creating demand, so protect their budget.

Once you have three months of data, a Looker Studio dashboard connecting GA4, Search Console and your sheet is worth building. Until then, the sheet is enough.

The "How did you hear about us?" field done right

This single form field is the highest-ROI attribution investment any lean team can make.

Use free text, not a dropdown. The expert consensus from Walker, Blend B2B, Medium Giant, and Cognism is unanimous. Dropdowns introduce selection bias: Ruler Analytics found that 47% of leads selecting the first option in a dropdown were inaccurately attributed. Free text captures granular insights like "My VP saw your founder's post about pipeline metrics on LinkedIn" or "Someone in the Exit Five Slack recommended you" - insights a dropdown of "Social Media" or "Community" would flatten into uselessness. Make it a multi-line text field to encourage detail, and make it required.

The conversion rate impact is negligible. Someone with genuine buying intent will not abandon a demo request form over one text field. Walker's approach is to remove one less useful field (like company size, or even company name, which can be enriched from other data sources) and replace it with HDYHAU rather than adding a net new field.

Place it only on high-intent forms. Enquiry, quote and demo requests, contact sales pages, pricing enquiries and free trial signups. Not content downloads, newsletter signups, or webinar registrations. Low-intent contacts haven't demonstrated purchase interest, and their attribution data is less strategically valuable.

Normalise responses in batches, not in real time. Wait until you have 30+ submissions before creating grouped categories. Common B2B categories include: LinkedIn, Podcast (with sub-categories by show name), Google/Search, Word of Mouth/Colleague, Community/Slack, Event/Conference, Blog/Content, and Review Site. Give AI assistants (ChatGPT, Perplexity, Gemini and similar) a category of their own. To sort a month's answers, paste them into an AI assistant with your category list; you don't need an automation for this.

Train sales to ask the deeper question (and make it easy for them to get it right).

The form captures the first layer. Sales discovery calls capture the second. The problem is that "How did you find us?" is the wrong question. It typically elicits a demand capture answer (Google, your website) rather than a demand creation answer (the thing that actually made them care).

Train whoever takes first calls to ask these two questions, in this order, on every discovery call:

1. "What first put us on your radar?" This surfaces the demand creation trigger. The answer might be a podcast episode, a post that landed in their feed, a colleague recommendation, or a panel session at a conference. It's almost never the Google search they used to find your website.
2. "What made you decide to reach out today?" This captures the conversion trigger: the specific moment or context that moved them from awareness to action. The answer is often a business event (new role, upcoming renewal, failed campaign) rather than anything marketing did.

If either answer is vague or surfaces "Google," follow up with: "Was there anything before that? Something you saw, or a recommendation?" A surprising number of buyers will then recall the actual trigger.

Two practical notes on making this stick. First, make logging the response a required CRM field to advance the deal stage and not a nice-to-have. Optional fields get skipped under pressure; required fields get filled in. Second, broadcast verbatim responses to a company Slack channel. When the founding team sees "I heard about you on [podcast name]" appearing alongside demo requests, the budget conversation about demand creation channels gets much easier.

Acknowledge the limitations honestly. Self-reported data has recency bias. Buyers tend to remember the last or most emotionally resonant touchpoint, not the first. They’ll also report the channel that feels most familiar or credible. A buyer who found you through a niche Slack community could report LinkedIn because it feels like a more legitimate answer. And be mindful of the Google problem: when someone says "Google," they often mean "I heard about you somewhere and then Googled your name". But, occasionally they genuinely did find you through search. The hybrid model accounts for this by treating software data (which shows the Google click) and self-reported data (which sometimes reveals the upstream trigger) as complementary layers, not competing ones.

Real-world evidence this works

Several B2B companies have documented their results with manual or hybrid attribution approaches.

Cognism, under CMO Alice de Courcy, shifted from lead gen to demand gen and added self-reported attribution to all high-intent forms. They discovered that dark social channels (LinkedIn organic, their podcast, and community mentions) were driving demand that software attribution had been crediting entirely to search. De Courcy documented the entire transformation in her ungated "Diary of a First-Time CMO."

Medium Giant, a B2B agency, tested self-reported attribution with a client heavily invested in content marketing over 12 months with 314 leads. Software-based attribution showed blog content and podcasts contributing almost nothing to pipeline. Self-reported data revealed they were the primary demand drivers. Conversion rates were unchanged after adding the “How did you hear about us?” field.

The pattern across these cases is consistent: teams that rely solely on software attribution systematically defund the channels that create demand (podcasts, Slack/Discord communities, LinkedIn thought leadership) because those channels don't generate trackable clicks. Hybrid attribution reveals the true picture and prevents this budget misallocation.

Your setup checklist

Work through it in order. Everything except the monthly review is one-off setup.

Set up once

• [ ] Set GA4 attribution lookback window to 90 days (Admin → Attribution Settings)
• [ ] Create UTM naming convention document in Google Sheets with approved source, medium, and campaign values
• [ ] Create "How did you hear about us?" multi-line text property in HubSpot
• [ ] Add HDYHAU as a required field on enquiry, demo, contact and pricing forms
• [ ] Set up Google Search Console branded query filter and record your baseline branded clicks, or use a branded Google Ads campaign if the filter isn't available for your site
• [ ] Create your four-column monthly tracking sheet: Month | Branded search (clicks) | Qualified enquiries | Revenue from enquiries (£)
• [ ] Schedule first Monday of each month: 60-minute strategic review
• [ ] Brief sales team on the two-question discovery sequence: "What first put us on your radar?" then "What made you decide to reach out today?" plus the Google follow-up probe
• [ ] Make sales-captured attribution a required CRM field to advance deal stage
• [ ] First Monday of each month: 60-minute strategic review. Compare self-reported vs software attribution, check branded search trend, review enquiries and revenue from enquiries
• [ ] Quarterly: check which channels buyers mention that the software barely credits, and protect their budget

Where this framework ends and paid tools begin

This system has real limitations worth being honest about. You won't get account-level journey stitching across multiple stakeholders in a buying committee. You won't get IP-to-company identification revealing which companies visit your website anonymously. You won't get automated audience syncing to ad platforms based on intent signals. And you won't get AI-driven analysis surfacing patterns across thousands of touchpoints.

If any of those capabilities would change your strategic decisions today, and you can justify £1,000+/month, look at Dreamdata (which has a legitimately useful free tier worth signing up for regardless) or HockeyStack. For companies spending under that threshold, Factors.ai offers a free plan identifying up to 200 companies per month with basic attribution, and Usermaven provides multi-touch attribution starting at around £80/year. Dreamdata's free tier includes basic account-level analytics and B2B benchmarks that complement this manual framework well. If you have a paid Hubspot Marketing plan, their Buyer Intent feature will surface account visits to your site.

But for the vast majority of growing businesses with marketing teams of one or two people, the constraint isn't data resolution; it's having any coherent measurement system at all. This framework gets you from zero to directionally accurate attribution within a month, for less than the cost of a team lunch. That's where you should start (but don’t compromise the team lunch).