outbound-unit-economics

Outbound Unit Economics

Overview

This skill turns an outbound plan into dollars per outcome: per lead, verified email, mobile, conversation, meeting booked and held, opportunity and deal. It prices data at ScraperCity catalog rates through the plan's credits, adds a dated tool stack, a loaded SDR cost built from published medians and the AE hours each meeting consumes, then runs the funnel with labeled benchmarks and tells the user what reply rate, win rate and cash the program needs.

The mistakes it prevents:

• Quoting data cost as the cost of a meeting. In the worked example data and tools are 4.8% of program cost; people are the rest.
• Pricing an SDR at salary. A $55K base SDR costs about $10,900 a month loaded once variable pay, employer cost, a manager share and ramp are in.
• Leaving AE time out of CAC, which hides a cost floor that no reply rate removes.
• Treating a 5-month CAC payback as a program that pays back in 5 months. Monthly cohorts keep the program cash negative far longer.
• Buying leads the mailboxes cannot send, or a sequencer tier that caps sends below mailbox capacity.
• Promising a cost per meeting from someone else's reply rate.

When to Use

• "What will cold email cost us per meeting?", "what's our CAC from outbound?", "cost per opportunity?"
• "What does an SDR really cost?", "is a half-time SDR enough?"
• "Which ScraperCity plan should we buy?", "Starter or Growth?"
• "How many leads and mailboxes do we need for 20 meetings a month?"
• "Does calling pay back for us?", "should we buy mobiles?"
• "How long until outbound pays for itself?", "how much cash do we need?"
• "Our actuals are below plan; where is the problem?"
• "Build me a model I can put in a spreadsheet."

Use something else for nearby jobs:

Prerequisites

Collect, or mark unknown and use the labeled default from references/unit-economics-model.md section 2:

1. Goal: meetings, opportunities or deals a month, and the decision the model must settle.
2. ScraperCity plan or plans to compare, and the data route (Lead Database, Local Business Database, Apollo URL, Maps, Store Leads, Email Finder).
3. Tool bills: mailbox seat, domain per year, sequencer tier, Sales Navigator, dialer, CRM seat share.
4. People: SDR share of time and pay, who takes first meetings and their pay, hours per held meeting and per opportunity.
5. Funnel actuals if any: valid, reply, positive share, meeting, held, opportunity and win rates.
6. Deal economics: ACV, gross margin, lifetime or monthly churn, billing (monthly or annual prepay), sales cycle.
7. Who on the team can own each action (founder, SDR, AE), so every recommendation gets an owner and a date.

When the user gives a total (for example "$7,000 a month for the SDR, tools and data") without the split, keep the total as the user input. Any split is an assumption, and no headline may rest on it: write the decision as a branch on the number they can look up ("if the SDR line is above $X loaded, cut hours; if below, keep the seat and add volume"), with both outcomes' CAC.

Authoritative Foundations

Workflow

1. Write the inputs table first

Every input gets a type: user input (their bills and salaries), catalog (ScraperCity price), benchmark (named source and year), assumption (a planning number the user replaces with a measured rate) or measured (their own data with a date range). Never put a benchmark in a user input row. A benchmark row names the report and its date; when the figure is computed from the source (a ratio of two medians, the 19% employer cost), the row shows that arithmetic so a reader can check it against the page. A figure built from several medians (a loaded SDR) is labeled as built, never as a reported number.

2. Price the data through the plan

catalog spend    = sum of (billable results x catalog price)
ScraperCity cash = plan fee + max(0, catalog spend - plan credits) / (1 + purchased-credit bonus)

Growth and Scale include the Lead Database and Local Business Database, where records cost nothing from the wallet within 100,000 new records a day. Starter and Growth tie at $149 of monthly catalog usage; Growth and Scale at $709. This model assumes unused credits do not carry over; check scrapercity.com/pricing.

3. Build the cost stack

List every monthly line with its basis, so the model knows what moves with meetings:

fixed      ScraperCity cash, mailbox seats, domains / 12, sequencer, Sales Navigator, SDR share
amortized  warmup months of mailbox and sequencer fees + setup, spread over 12 months
variable   AE hours per held meeting and per opportunity x AE loaded hour
SDR loaded per productive FTE = ((base + variable x share paid) x 1.19 / 12 + manager OTE x 1.19 / 12 / SDRs per manager)
                                x tenure / (tenure - ramp / 2)
AE loaded hour = AE OTE x 1.19 / 2,080

Defaults and dated tool prices are in references/cost-stack.md. Allocate shared people by hours and shared seats by use.

4. Check capacity before sizing the list

mailboxes         = domains x mailboxes per domain
sends per month   = mailboxes x sends per mailbox per day x business days
contacts capacity = sends per month / touches per contact
contacts emailed  = min(valid emails, contacts capacity)

The sequencer tier caps both: Instantly Growth ($47) allows 5,000 emails and 1,000 uploaded contacts a month, Hypergrowth ($97) 125,000 and 25,000 (instantly.ai/pricing, checked 2026-10-13). So contacts emailed = min(valid emails, contacts capacity, tier contact cap, tier email cap / touches). Compare held meetings per SDR FTE with Bridge's median quota of 10.

When the user knows meetings booked but not contacts emailed, start from their counts. The top of the funnel is then a diagnostic with two branches (full capacity and an implied low reply rate, or a capped list and idle mailboxes), each with its implied reply rate and its fix, and the first measurement that picks the branch. Neither branch is the base case, and no headline rests on one. SDR utilization comes from the SDR's logged hours, never from a benchmark activity rate.

5. Run the funnel

replies = contacts x reply rate            positives = replies x positive share
meetings = positives x meeting rate        held = meetings x held rate
opportunities = held x opportunity rate    deals = opportunities x win rate

For phone: conversations = dials x connect rate x conversation rate; meetings = conversations x meeting rate (cold-calling).

6. Compute unit costs

7. Test against break-even

monthly gross margin per deal = ACV / 12 x gross margin
CAC target  = margin x 6     CAC ceiling = margin x 12     payback = CAC / margin
LTV to CAC  = margin x lifetime months / CAC        (lifetime = 1 / monthly churn when churn is known)
AE cost per deal (CAC floor) = AE per held / (opportunity rate x win rate) + AE per opportunity / win rate
highest cost per booked meeting = CAC limit x held rate x opportunity rate x win rate
break-even reply rate = fixed / (contacts x (k_deal x CAC limit - k_held x AE per held - k_opp x AE per opportunity))
   where k_held = positive share x meeting rate x held rate, k_opp = k_held x opportunity rate, k_deal = k_opp x win rate

Then run the cash curve (references/unit-economics-model.md section 6): month 0 warmup, deals sign after the sales cycle and pay monthly or a year up front. Report the break-even month and the lowest cumulative cash.

8. Run sensitivity

One grid: cost per deal across at least five reply rates (1.5%, 2.5%, the benchmark, Hunter's 4.5%, 5.5%) by four win rates, every cell with its own total, since AE time moves with meetings. One scenario table: one lever at a time against today, each row a full re-run (unit-economics-model.py scenarios levers.json) so total, meetings, deals, cost per meeting, CAC and payback agree. Name the stage that moves CAC most. No second grid.

9. Size to the goal and recommend

Reverse the funnel for the user's target: contacts = target meetings / (reply x positive share x meeting rate), leads = contacts / valid rate, mailboxes = contacts x touches / (sends per day x business days), domains rounded up. State the plan, list size, mailboxes, reply and win rates to beat, cash to fund and the first measurements. scripts/unit-economics-model.py target N does the arithmetic.

10. Write the fixes, with owners and copy

Each fix the recommendation names comes with a named source for the tactic, an owner who exists on the user's team, a calendar date and every message it involves written out in full. Stage fixes and their copy are in references/break-even-and-sensitivity.md section 12: Reisert's confirmation sequence and no-show rebooking for the held rate, a Pessar-style qualifying question before the calendar link for the opportunity rate, and an SDR quota counted in held meetings and converted opportunities as Bridge reports them. When deals a month are few, show deals a quarter and the quarterly CAC at one deal fewer and one more, so the user judges CAC on enough deals.

11. Replace assumptions with actuals

As each stage produces measured counts, swap its assumption for the measured rate, with the sample size and date range beside it, and re-run. When actuals miss plan, split the gap by stage (unit-economics-model.py variance actual.json) and fix the stage with the largest share first.

ScraperCity Implementation

The model runs on catalog prices; use tool runs only to measure rates cheaply:

Data cost per verified email by route, at a 70% valid rate (an assumption; measure yours with the 200-email test):

Worked cost estimate (data line only)

Starter is $93.84 a month cheaper at this volume, 0.9% of the worked example's $10,363.14 program cost. Growth buys same-day database pulls, excludeDelivered dedupe and no multi-day Apollo delivery. Spend over $5 still needs a check_wallet call and the user's go-ahead before any run.

Output Format

One Markdown model from templates/output-template.md: a one-line context, the answer, then Inputs, ScraperCity Cost, Cost Stack, Capacity, Funnel, Unit Costs, Break-even, Sensitivity, Recommendation and Sources, and no other sections. Show arithmetic in cells as a x b = c. Compute every figure with scripts/unit-economics-model.py (or the spreadsheet checked against it) and run the checker before sending.

• Answer first. After the context line, at most five sentences: the decision, CAC and payback, the levers that move them, each number once. Anything that rests on an assumption is an if/then with both numbers. One line then says which figures are the user's; a "nothing was run" note appears there once or not at all.
• Each fact once. No second Cost Stack table (the SDR and AE derivation is one line under it), one grid plus one scenario table, and no row that repeats a figure another section states. At most 2,000 words.
• Only what was asked. ScraperCity Cost carries the plan row, the catalog lines and the cash row. Plan comparisons, unused-credit ideas and route tables appear only when the user asked which plan to buy; otherwise one line ("data line at $149 on Growth; anything you pay above that is a direct cut"). No frameworks, scores or ratios the user cannot act on.
• Plain words. Write "a count of matching leads" and "validate 200 current emails" in place of tool ids, parameter names, code spans, file paths, skill or script names. Leave out handoff lines to other skills, disclaimers and self-graded checklists.
• Recommendation table columns: Decision, Value, Why, Owner, Due. It answers the question first, then names the reply rate and win rate to beat, the cash to fund and the first measurements. Owners are roles the user has; due dates are calendar dates.
• Every message the plan names is written out below the table as a quote block, ready to send: confirmation, reminder, no-show, qualifying question.
• Sources lists each external figure's source name, date and URL, one line each.
• When the user asks a narrow question (only "Starter or Growth?", only "what does an SDR cost?"), answer it in a few lines with the formula and number and offer the full model.

Quality Check

python3 scripts/check-output.py model.md enforces:

• The nine sections in order plus Sources and nothing else, every [fill] replaced, 500 to 2,000 words, at least 30 table rows; legal content, dashes, file paths, code spans, tool ids and parameters, skill names, disclaimers and a repeated "nothing was run" note fail.
• Every Inputs row is typed and every benchmark names a year; benchmarks require a Sources section with URLs.
• ScraperCity lines use catalog prices by product name with count x price = cost; the plan row shows its fee and credits; the cash line equals fee + max(0, spend - credits) / (1 + bonus).
• One Cost Stack table whose rows are count x unit cost and sum to the total, carrying the ScraperCity cash; per-meeting and per-opportunity lines use the Funnel's counts.
• Each Funnel count equals the previous count times its rate; cost per meeting booked, held, opportunity and deal each equal total / count within 1%.
• Every a x b = c cell and Formula and Value pair computes, with at least 15 shown.
• Payback = CAC / monthly margin; the ceiling is 12 months of margin; the 12 month break-even reply rate is recomputed.
• Every cost per deal grid cell is recomputed; every scenario row has CAC = total / deals, cost per meeting = total / meetings and payback = CAC / margin.
• Any line sizing Instantly Growth states its 1,000 contact cap.
• The Recommendation has Owner and Due (a calendar date) on every row, names the reply rate to beat, and writes out as many messages as it names.

Also check by reading: no headline rests on an assumed split or volume; the cash curve and CAC payback are both stated when the user asks about payback; nothing presents a benchmark as a promise.

Common Pitfalls

Execution Artifacts

Related Skills

• cold-email-strategy designs the cells this model prices.
• cold-calling and multi-channel-outreach supply the phone and multichannel funnels.
• signal-triggered-outbound adds repeat-pull costs and higher-converting tiers.
• domain-infrastructure, inbox-setup and sending-platforms set the capacity and tool costs.
• email-deliverability protects the reply rate the model depends on.
• reply-handling lifts the positive share and meeting rate.
• icp-list-builder, mobile-appender and scrapercity-lead-database produce the leads and mobiles priced here.