The True Cost Per Lead of a Multi-Channel Prospecting Stack (With a Worksheet)

What actually goes into your cost per lead across channels?

Most teams read their cost per lead off the scraper invoice and stop there. That number is fiction. The lead that lands in a rep's inbox has passed through five or six paid tools, and each one took its cut on the way.

A typical stitched-together stack for B2B outbound looks like this:

• A scraper or data provider to pull companies from Google Maps and LinkedIn
• An enrichment tool to fill the missing phone and email, because the scrape comes back half-blank
• An email platform with its own per-seat or per-send pricing
• A WhatsApp tool, usually a separate vendor charging per message
• A dialer or voice tool for the calls
• CRM seats so reps can actually work the records

Six subscriptions, six renewal dates, six pricing models that never talk to each other. Three bill per seat whether the seat is busy or idle. Two bill per credit that expires at month-end. One bills per message with a minimum commit. None of them collapse cleanly into a single number you can put in front of a CFO.

The real formula is short, and it stings:

• Cost per worked lead = (every monthly tool fee + every per-unit charge) / (leads you actually contacted, not leads you scraped)

The denominator is where the damage hides. You scrape 5,000 records, but 1,800 are duplicates or dead, enrichment fills only part of the rest, and reps end up working maybe 2,400. You paid for 5,000. You divide by 2,400. The sticker price doubles before a single message goes out. That gap between scraped and worked is the line nobody puts on the spreadsheet, and it is the biggest single reason teams lowball what prospecting really costs.

How much does a stitched stack cost versus one metered platform?

Let's run one honest scenario. A mid-size outbound team: six operators, targeting 4,000 worked leads a month split across Google Maps and LinkedIn, with email and WhatsApp as the main outreach channels and some calling on top.

The stitched stack, conservative monthly numbers:

• Scraper / data provider: $99 base, plus credits
• Enrichment tool: $150 at this volume
• Email platform: 6 seats at roughly $40 = $240
• WhatsApp tool: $80 base plus per-message
• Dialer: $90 base plus per-minute
• CRM: 6 seats at $50 = $300

That is about $959 of subscription floor before a single lead moves, and the per-unit charges for enrichment credits, messages, and minutes stack on top. Call it $1,200 to $1,400 all-in for the month. Divide by 4,000 worked leads and you land at $0.30 to $0.35 per worked lead, and that still ignores the dead records you paid to scrape. Most of that money is seat rent for tools sitting idle two weeks out of four.

Now the metered alternative, priced at ArivonHub's published unit economics:

• Google Maps leads: $0.009 each
• LinkedIn leads: $0.029 each
• WhatsApp AI messages: about $0.004 each
• Email sends: $0.0009 each

Say 2,500 Google Maps leads and 1,500 LinkedIn leads for the month: $22.50 plus $43.50 = $66 in lead collection. Add 8,000 WhatsApp messages at about $0.004 = $32, and 20,000 email sends at $0.0009 = $18. Collection plus outreach comes to roughly $116 for that volume, on top of one plan that already covers the channels, the CRM, and the AI in a single subscription. The Growth plan, at $299 a month, includes 5,000 Google Maps and 5,000 LinkedIn leads plus WhatsApp and AI replies before you buy a single pack. No per-seat tax on the operators working the queue, and no second renewal date to reconcile.

The two are not in the same weight class. One charges you for capacity you reserved; the other charges you for leads you actually pulled and messages you actually sent. When volume dips in a slow month, the stitched stack still bills the full $959 floor. The metered platform bills less, because you used less. That asymmetry is the whole argument.

Why do seat-based and credit-based tools quietly inflate cost?

Two pricing mechanics do most of the damage, and both are built to look harmless on the order form.

The first is the per-seat tax. You buy six CRM seats and six email-tool seats because you have six reps. But prospecting is spiky. Two reps are on leave, one is ramping, and your real heavy-usage seat count this month is three. You still pay for twelve. Seat pricing charges you for the org chart, not the work. It punishes you for hiring ahead of demand, then punishes you again when someone leaves mid-cycle and the seat sits paid-but-empty until renewal.

The second is the expiring credit. Enrichment and data tools love to sell credits in monthly blocks that reset. Buy 5,000, use 3,100, lose 1,900. That is not a discount, it is forfeiture. You pre-pay for a forecast and eat the variance every single month. Over a year, the forfeited credits alone can equal a full extra month of spend.

Metered consumable packs flip both problems. On ArivonHub the packs run $9 for 1,000 Google Maps leads, $12 for 1,000 Facebook or Instagram leads, $29 for 1,000 LinkedIn leads, $19 for 5,000 WhatsApp AI messages, and $9 for 10,000 email sends. You buy a pack when you need capacity, the credit sits in your ledger, and there is no month-end clock forcing you to burn it or lose it. A slow month costs less. A heavy month, you top up. The spend tracks the work instead of the calendar.

This matters most for the objection every operator raises: will the pricing surprise me with overages? On a stitched stack, overage is structural. You blow past a credit tier mid-month and get bumped to the next bracket, or you hit a send limit and pay a penalty rate. With fixed-price packs there is no overage mechanic at all, because you buy capacity in known blocks before you use it. What you see on the pack is what you pay. No bracket creep, no surprise line on the invoice.

How do I run this worksheet on my own numbers?

Here is the worksheet. Pull your last full month of invoices and fill in the blanks. It takes about fifteen minutes and it usually changes the conversation with finance.

Step 1 — Total your fixed subscription floor. Add every base subscription that bills whether you use it or not: scraper base, enrichment base, email platform, WhatsApp base, dialer base, CRM seats. This is what you pay even in a dead month.

Step 2 — Total your variable per-unit charges. Add enrichment credits consumed, message fees, dialer minutes, and any send overages from last month. Be honest about overages; that is where the surprises hide.

Step 3 — Find your real worked-lead count. Not scraped. Worked. The records a rep actually touched with an email, a message, or a call. Pull this from your CRM activity log, not your scrape export.

Step 4 — Divide. (Step 1 + Step 2) / Step 3 = your true cost per worked lead. Write it down. This is the number that was hiding.

Step 5 — Model the metered version. Take your channel mix and price it at the published units: Google Maps leads at $0.009, LinkedIn at $0.029, WhatsApp at about $0.004, email at $0.0009. Add the plan that covers your channel count and AI needs. Divide by the same worked-lead count.

Step 6 — Compare the two cost-per-lead numbers, then look at the floor. The metered number is usually lower, but the bigger insight is the floor: in a month where your volume drops by half, the stitched stack barely moves while the metered spend roughly halves. That downside protection is worth more than the headline price difference, because outbound volume is never flat.

Two traps to avoid. First, do not divide by scraped leads to flatter your current stack; use worked leads, or the comparison is dishonest in your own favor and you will regret it next quarter. Second, subtract the engineering and ops time spent reconciling six invoices and re-syncing data between tools that were never built to talk to each other. That time is real cost; it just never shows up on any single invoice. A team that consolidates onto one queue gets that hour back too.

Does a lower cost per lead mean weaker quality or less control?

This is the fair pushback. A low cost per lead is worthless if the leads are junk, or if consolidating onto one platform means handing control to an autonomous black box. Neither is the trade here.

Quality is handled before the lead ever reaches a rep. AI enrichment fills the missing phone, email, and company data, scores every lead for readiness, and flags duplicates, so reps work qualified records instead of half-blank rows. That is the lever that actually moves your cost per worked lead, because it shrinks the gap between scraped and worked. A cheaper unit price on a record you would have thrown out anyway is not a saving. A qualified record at $0.009 is.

Control is the part operators worry about most, and the part most worth spelling out. The AI here drafts and suggests; it does not run unsupervised. Auto-reply sits behind real guardrails: manual takeover stops it instantly, a cooldown keeps it from over-firing, quotas cap volume, and plan and tenant toggles can switch it off entirely. The operator owns review, escalation, and the final word on what goes out. That is the opposite of a passive CRM that just stores records and waits, and also the opposite of an autonomous tool that messages your prospects while you are not looking. AI accelerates the busywork; humans stay in command of the relationship.

The consolidation argument and the quality argument reinforce each other. One queue means one place where enrichment, scoring, outreach, and oversight live together, so a lead flagged as low-readiness never burns a paid message, and a conversation that needs a human gets one before the AI talks past it. On a stitched stack those signals are stranded in separate tools and the handoffs leak. Lower cost per lead and tighter control are not in tension here. The same architecture that meters your spend is the one that keeps a human in the loop.

If you want to test it against your own numbers before committing, the free 14-day trial gives you Google Maps lead collection with no credit card, which is enough to validate the worked-lead math on real data instead of a spreadsheet estimate.