Marketing attribution used to feel like detective work. Somebody clicked an ad, visited your website, filled out a form, joined your email list, booked a call, and eventually bought something. You could follow the footprints and feel reasonably confident about what worked.
Now the footprints look like Bigfoot walked through a server farm wearing roller skates. In this episode of the Human-First AI Marketing Podcast, I talked with Keith Perhac, founder of SegMetrics, about marketing attribution, AI analytics, zero-click marketing, and the very strange future of knowing where your leads actually came from.
Keith shared a stat that should make every SMB owner and marketing director sit up a little straighter: more than half of internet traffic may now be bots, scrapers, AI agents, and other non-human activity. He also pointed out that fewer searches are turning into website visits because people are getting answers directly from Google, AI tools, YouTube, social platforms, and other walled gardens.
That means your buyer may know a lot about you before they ever show up in your analytics. Your website used to be the “main office” of your digital presence. Every ad, email, social post, and referral tried to bring people there. Now your brand has become more like a franchise with locations everywhere: your website, LinkedIn profile, YouTube videos, podcast interviews, review sites, social posts, Reddit mentions, and AI summaries all become part of the buyer’s research journey.
Keith described this shift clearly. AI is taking information from your website, social profiles, YouTube, Capterra, and what other people say about you, then creating a new version of your “website” inside the AI answer. That shift matters because most SMBs are still measuring marketing like every serious buyer walks through the front door. Many of them are circling the building, asking neighbors, reading reviews, watching a video, checking your LinkedIn profile, and asking ChatGPT if you are worth calling.
The Attribution Problem Is Really a Human Understanding Problem
Keith made one of the most useful points of the whole conversation: attribution means more than knowing where people came from. It means understanding who they are, what they do, what they need, and which groups are taking the actions you want them to take.
That is a huge mindset shift for SMBs. A lot of marketing teams obsess over the source field in HubSpot, Google Analytics, or their ad dashboard: paid search, organic social, referral, direct, email. Those labels can help, but they rarely tell the whole story.
For B2B companies, especially small and mid-sized companies with lower traffic volume and longer sales cycles, the better question is often: which types of people are converting? A manufacturer, franchise, consultant, SaaS company, or professional services firm may never have enough website traffic to make perfect decisions based on button colors, micro-clicks, or tiny A/B tests. Keith and I talked about how segment analysis can matter more than broad metrics for B2B businesses because each good-fit person on your list may be genuinely valuable.
Think of your analytics like a farmer’s market. A big consumer brand may be running a massive grocery store with thousands of people coming through the aisles every hour. They can study traffic patterns, shelf placement, coupon behavior, and checkout speed with statistical confidence.
Most SMBs are working a smaller booth with better conversations. You may only talk to 20 people, but three of them might be perfect buyers. Your job is to notice who leaned in, what they asked, what they bought, what they ignored, and what brought them to your table in the first place. That is where human-first analytics begins.
Ask People. Then Save What They Tell You.
One of Keith’s best practical recommendations was beautifully simple: ask people who they are and save that information in your CRM. That sounds almost too obvious, which is how you know marketers will try to avoid it and buy six tools instead.
Keith explained that SegMetrics asks people during onboarding whether they are in e-commerce, coaching, agency work, creator businesses, or other categories. Then they can look at how those different groups move through the funnel, what they convert on, what objections they have, and what plans should be built around that information.
This is first-party data with a pulse. Relying only on ad platforms to guess who someone is will leave gaps in your decision-making. Ask directly. Add the answer to a contact record. Use tags or custom fields. Connect that information to actual behavior and revenue.
For SMBs, this can start with a few simple questions:
- What industry are you in?
- What role do you play in the buying decision?
- What problem are you trying to solve?
- How soon do you need help?
- What triggered your search?
You need a CRM, a few smart fields, and the discipline to capture the same information consistently. That small habit can become one of the most useful parts of your marketing system because it connects the human story to the performance data.
Use AI to Find the Outliers
Keith’s most practical AI analytics advice was to use AI to find the biggest conversion modifiers in your funnel. In the episode, he described a SegMetrics onboarding funnel with a 50% average conversion rate. That number sounds useful, but averages hide the story. When AI analyzed the data by industry and source, it found that some segments converted extremely well while others lagged far behind.
That is where the money is. Averages are the beige carpeting of analytics. They cover everything and reveal very little.
Outliers give you choices. When one segment converts at 80% and another converts at 20%, you can make a real business decision. You can build a better funnel for the underperforming group, or you can decide that group is outside your ideal market and stop spending energy there. Keith made the point that both choices are legitimate. The important part is making the choice intentionally.
This is a great use of AI because it gives the machine a job it is actually good at: looking through structured data and finding patterns faster than a tired marketer staring at a spreadsheet after lunch. AI can look across industries, lead sources, company sizes, campaigns, personas, funnel stages, custom fields, and outcomes. Then humans decide what matters.
That is the Human-First AI Marketing® pattern we believe in at Avenue9: human strategy first, AI leverage second, better decisions always.
Take One Funnel at a Time
Here is where Keith saved every marketing director from building a 49-part funnel monster. When I asked about multiple industries, personas, and buyer journey stages, Keith recommended taking funnels one at a time. Start with a real funnel. Ask AI to find the biggest outliers. Improve the most meaningful gap. Let it run. Learn from it. Then move forward.
This matters because AI makes overbuilding very tempting. You can create 12 nurture sequences, 9 personas, 47 landing pages, and 300 social variations before your coffee gets cold. That does make you feel productive. It also gives your team a beautiful new maze to get lost in.
Start with the fruit stand before you build the grocery empire. Pick one offer, one funnel, one audience segment, and one major leak. Fix that first. SMB marketing gets stronger when the system stays simple enough for a real team to run.
AI Should Turn Analytics Into Decisions
Keith also made an important point near the end of the episode. Marketers already have too much information. AI can make that worse by giving us pages and pages of analysis, summaries, charts, and recommendations. More data can quickly become another swamp.
The value comes from decision clarity. A good AI analytics workflow should help you answer practical questions that change your next move:
- Where are the best-fit leads coming from?
- Which segment converts best?
- Which segment needs a different message?
- Which funnel step is creating friction?
- Which campaign deserves more investment?
- Which activity can be retired?
This is why Keith described SegMetrics’ AI MCP server as a way to connect AI directly to marketing analytics so it can query real data rather than guess from a CSV dump. That direction is exciting because AI analytics can become more like a good analyst sitting beside you, helping you find the signal, challenge assumptions, and decide what to do next.
The Human-First Lesson
The biggest takeaway from my conversation with Keith is simple: marketing analytics should help you understand people better. Bots can inflate traffic, AI answers can hide buyer journeys, ad platforms can guess wrong, and dashboards can overwhelm. Your best path forward is to collect better first-party data, segment your audience thoughtfully, use AI to find meaningful patterns, and keep human judgment in charge of the decision.
At Avenue9, we believe AI should amplify your authenticity and help your team make clearer, faster, more human decisions. Marketing attribution is changing fast, and SMBs have a real opportunity here. Big brands may have more data, but smaller companies often have better customer conversations, closer sales feedback, and clearer founder stories. That is your advantage.
Start by asking better questions. Save the answers. Analyze the patterns. Improve one funnel at a time. When the analytics start looking like Bigfoot roller-skated through your dashboard, bring in a human-first guide.
Avenue9 helps small and mid-sized businesses turn real expertise, customer conversations, and first-party insights into AI-powered marketing systems that build trust, improve execution, and support better growth decisions. Visit Avenue9.com to learn how our Human-First AI Marketing® approach and A9^Factor Framework™ can help you make sense of AI, attribution, and your next best marketing move.
