An Outbound Agency Founder Published His Actual Funnel. We Checked Every Stage.

Every agency founder eventually publishes a funnel breakdown. It's a genre at this point — a tidy three-stage diagram, a handful of tool logos, a headline number nobody outside the building can actually check. I don't say that to dismiss the genre. I say it because it's true, and the honest way to read one of these posts is to take it seriously and check it hard at the same time. Not one or the other.

Michel Lieben, who runs ColdIQ, a cold outbound agency, published exactly this kind of post on July 31: "The Funnel That Books Us 100+ Meetings Every Month." It has real numbers in it — traffic sources, named tools, a direct quote about how deals actually close. So I did what I'd want someone to do with a funnel breakdown of my own: read it stage by stage, and ask what has to be operationally true for each claim to hold at the volume he's describing.

The Funnel This Cold Email Agency Says Books 100+ Meetings a Month

Start with the traffic, because it's the part of the post you can actually check against itself. ColdIQ gets roughly 40,000 monthly visitors. Lieben breaks that into four sources: about 12,000 from search (SEO and AEO — standard articles, meta pages, programmatic SEO pages, and free mini-tools), about 10,000 from LinkedIn (his team runs 10-plus people posting roughly 100 times a month combined), about 3,000 as a byproduct of running cold outbound campaigns themselves, and the remainder — a little over a third of total traffic — from YouTube and recent webinars, some of which pulled 2,000 to 4,000 registrants apiece.

Worth saying plainly, once: none of this is independently verified. It's Lieben's own dashboard, published on his own blog, with no outside audit underneath it. That's not a reason to wave it off — the specificity of the traffic breakdown, the named tools, and the fact that ColdIQ has published under its own name for years all count for something real. It's just the correct caveat to state before anyone builds a strategy around someone else's self-reported funnel, and it's the same posture we brought to another agency's self-reported outbound claim worth reading against this one.

Stage One: Four Lead-Capture Mechanisms Running at Once

Stage one is where ColdIQ tries to catch all 40,000 of those monthly visitors before they leave without a trace. Lieben names four mechanisms, and he's specific that they run concurrently, not as a sequence a visitor moves through one at a time. Contextual lead magnets that change based on what someone is actually browsing. Email nurture sequences once an address is captured. Newsletter and e-book calls-to-action seeded across the site. And visitor de-anonymization, paired with LLM-based lead scoring and segmentation, that identifies who showed up even if they never filled out a form.

Running four capture mechanisms at once instead of picking one is a reasonable design choice, and it matches the traffic mix — a visitor arriving from a search result behaves differently than one arriving from a LinkedIn post, and a single capture method would miss most of them. But four concurrent mechanisms also means four separate streams of leads landing somewhere, at some volume, needing some kind of triage before anyone can call them a qualified lead. That's the part stage one hands off to stage two, and it's also the part the post spends the least time explaining.

Stage Two: What "70% Sales-Led" Actually Requires at This Volume

“70% of our deals are sales-led. They close through conversations on Google Meet, usually across one or two calls.”

Read on its own, that's a clean, confident line. Read against the traffic numbers above it, it raises the actual question worth asking: who is triaging 40,000 monthly visitors, four concurrent capture mechanisms, and 10-plus people's worth of LinkedIn-driven inbound down to a number of qualified conversations small enough for a handful of reps to run one or two calls each and close 70% of the resulting deals?

The de-anonymization and LLM scoring from stage one is the closest thing to an answer the post gives, and it's a real answer as far as it goes — automated scoring absolutely can narrow 40,000 visitors down to a manageable shortlist. But scoring isn't the same operation as booking and qualifying an actual meeting. Somebody, or some team, still has to take that scored list and turn it into calendar invites, and the post names the tools that show up after that point — Claap for recordings, Cobl for the sales workflow around them — without saying much about the qualification step that sits right before a lead becomes a call. That's the least documented stage of the three, and it's also the one carrying the most weight behind the 70% number.

None of that makes the claim false. A 10-plus person team that's disciplined about lead scoring can absolutely route inbound efficiently enough to make 70% sales-led plausible at this scale. But "plausible if the triage layer is well-built" and "true because it's stated" are different levels of confidence, and the post only earns you the first one. If you're trying to reproduce this funnel rather than just admire it, the qualification step between de-anonymization and the Google Meet invite is the piece you'd need to build and staff yourself — it's not handed to you in the post.

Stage Three: What This Cold Email Agency's Funnel Doesn't Mention

Stage three is the part of the post with the clearest mechanics: call recordings and transcripts feed proposal generation and sales coaching, using Claap and Cobl. That's a sensible, well-worn pattern — record the call, mine the transcript, generate the proposal off what was actually said instead of what a rep remembers saying. Nothing to argue with there.

What's conspicuously absent from the whole post is any mention of what it costs to produce that ~3,000 monthly visitors ColdIQ gets "as a byproduct" of running cold outbound campaigns. That phrasing makes it sound incidental, almost free — traffic that shows up while you're doing something else. It isn't free. Generating even a byproduct of that size means running real send volume, which means multiple sending domains, a warmup schedule, verified lists, and ongoing deliverability monitoring, all before a single one of those 3,000 visits shows up in an analytics dashboard. We ran the actual unit economics behind that kind of volume in our own audit of cold email agency math, and the sending infrastructure line item is exactly the one that tends to go unmentioned in a founder's funnel post, for the same reason it goes unmentioned here.

It's also worth noting this isn't the first time this specific company has published something worth reading with one eyebrow raised. We covered ColdIQ's own pivot away from campaign teardowns toward b2b data buyer's guides a few days before this post went up — a different kind of confident, self-reported claim from the same agency, in the same window. That's not a strike against Lieben. It's a reason to read both posts with the same posture: real information, published in the founder's own interest, worth taking at face value and checking anyway. Anyone running outbound in the broader B2B space should expect more of this genre, not less, as agencies compete on published proof instead of just claimed results.

The Real Lesson: Build a Funnel That Survives Losing One Channel

Here's the part of Lieben's post that doesn't need independent verification to be useful, because it's a structural claim, not a performance claim. He says the three-stage architecture — capture, conversion, closing — has stayed the same for three years, even while the individual tools underneath it rotated repeatedly. Prospeo, GetLeads, PredictLeads, and FullEnrich for data and enrichment. Expandi, lemlist, Instantly, and Apollo for outreach. Taplio, beehiiv, and Customer.io for content and nurture. Claap and Cobl for sales. That's a lot of tool churn sitting underneath a funnel shape that never changed.

The argument underneath that stability is the one worth taking home, and it's the diversification point: four traffic sources means losing any single one just reduces volume, it doesn't collapse the funnel. Lose the LinkedIn channel and ColdIQ still has search, cold outbound's own byproduct, and YouTube plus webinars. Lose a tool in the outreach stack and the architecture absorbs it, because the architecture was never built around one tool in the first place. That's a genuinely portable idea, and it doesn't require ColdIQ's scale to be true. A funnel with two traffic sources instead of four is one bad platform update away from a real problem. A funnel with four is just thinner for a while.

So take the numbers with the caveat they deserve — self-reported, unverified, plausible rather than proven at the stage-two triage layer, and quiet about the deliverability cost sitting under that cold-outbound byproduct traffic. But take the structural lesson at full value: the architecture matters more than which tools sit inside it, and the number of independent traffic sources feeding your pipeline is a real resilience metric, not a vanity one. Count your own traffic sources this week. If losing your best one would zero out your pipeline instead of just shrinking it, that's the fix worth making before you copy anything else out of this post. If you'd rather have someone build that resilience into your outbound program directly, that's what our cold email team does before we write a single sequence.

See where you are cited today

A free snapshot audit of your rankings and AI citations before we ever talk.