The agency-prospecting glossary

Close Probability Score

A statistical estimate of how likely a specific prospect is to close into a paying client, trained on the agency's historical outcome data — the long-run replacement for heuristic lead scoring.

Close probability scoring is the natural endgame of an agency's lead-scoring journey. Once the agency has logged 50-200 outcomes per vertical (closed-won, closed-lost, no-show, ghosted), a statistical model can predict close probability for new prospects with materially better accuracy than a fixed-weight heuristic.

The catch: the data quality is everything. The agency needs accurate outcome tracking back to the original prospect data — most CRMs lose this link the moment the prospect becomes a contact. Tools that maintain the bridge between the prospect signals (at the time the agency contacted them) and the eventual outcome are rare and expensive to build.

For agencies with under 50 closed-won outcomes per vertical, close probability scoring is premature — the model has nothing to train against and reverts to the underlying heuristic anyway. For agencies past that threshold, the lift over heuristic scoring is typically 20-40% on close-rate per dial.

Market Opportunity Score

A 0-100 score grading an entire city or zip code on agency-prospecting attractiveness — derived from demographics, business density, and historical close-rate signals.

Where an opportunity score grades one business, a market opportunity score grades a whole geography. Inputs include median household income, home-ownership rate, owner-occupied housing units, median home value, total housing units, median age, and density of the local-services verticals an agency typically targets.

Markets with high MOS scores tend to share three traits: high enough income that local-services businesses can afford a $1k-3k/month retainer, enough housing density that the agency's customer's customers exist, and enough business diversity that the agency's pitch isn't fighting a saturated competitor pool. Working-class metros with thin business density score lower; high-income suburbs with diversified local-services economies score highest.

Practically, MOS lets agency owners pick which markets to scan first. An agency with 10 free scans a month should spend them on MOS 70+ markets where the underlying economics support agency retainers, not on MOS 30 markets where every prospect is going to push back on the price.

Opportunity Score

A 0-100 composite ranking how marketing-ready a local business is, derived from review volume, web presence, ad activity, and demographic context.

An opportunity score (sometimes called a website opportunity score, agency opportunity score, or vulnerability score depending on the tool) compresses 15-20 underlying signals about a local business into a single 0-100 number. The higher the score, the more visible signs of under-investment in marketing — and the better the agency-pitch fit.

Typical inputs: review count and recency, star rating, whether the business has a website, how recently the website changed, presence of ad-tracking pixels, social-media presence, hiring signals, and per-capita competitor density in the surrounding market. Some scores layer demographic context to weight markets where customers can actually afford the agency's typical retainer.

Used well, an opportunity score replaces gut-feel triage. An agency owner with 300 prospects on a Friday afternoon can rank by score, call the top 20, and skip the bottom 100 with a calibrated sense of what they're skipping. Over weeks, the agency learns which score band fits their offer best — and tightens outbound targeting accordingly.

Agency Prospecting

The full process by which a marketing agency finds, qualifies, and converts local businesses into paying clients — distinct from B2B SaaS sales prospecting in tooling, signals, and conversion expectations.

Agency prospecting at the local-services level is a different discipline from enterprise B2B sales prospecting. The buyer is the operator (plumber, dentist, roofer, salon owner), not a marketing director. The pitch is concrete (a website, ads, SEO) not abstract (digital transformation, ROI uplift). The signals that matter are visible-on-Google-Maps gaps in the prospect's marketing stack — not intent data or buying-committee composition.

A typical agency prospecting workflow: pull a list of businesses in one market and one vertical, filter for buying signals (no website, bad reviews, no ads pixel, sub-4-star rating), score each prospect, run DNC and line-type scrubs, export the survivors to a CRM or outreach tool, run sequences sized to the agency's offer, log outcomes back to the system.

The whole loop typically takes 30-60 minutes per market when the tooling is right, and 4-8 hours when done manually in a spreadsheet. The breakeven on a paid prospecting tool is one closed retainer per quarter.

Marketing agency prospecting — the full playbook & tool →

Cold Outbound Cadence

The structured sequence of touches (call, email, voicemail, SMS, follow-up) an agency runs against each cold prospect — the difference between random outreach and a repeatable conversion engine.

A cadence is the choreography of cold outbound. A typical local-services agency cadence: Day 1 call + voicemail, Day 1 email, Day 3 second call, Day 5 LinkedIn touch (where relevant), Day 8 third email referencing the prior touches. Specifics vary; the discipline of running the same cadence on every prospect is what produces measurable conversion data.

Without a cadence, the same prospect gets contacted twice in three days by mistake, or once and then forgotten. Replies don't compound. The agency owner's brain has to track what happened with each prospect manually — and when they get to prospect #50, the early prospects fall off the radar.

Tools enforce cadence by holding the schedule (sequence platform, dialer with scheduling) and the agency owner just executes the next-step queue each morning. The point is not the perfect cadence — it's the same cadence, on every prospect, for long enough to learn what's working.

Lead Scoring

The discipline of ranking prospects by likelihood-to-buy before spending outbound effort on them — produces a prioritized call sheet instead of a flat list.

Lead scoring takes the underlying signals about a prospect — for an agency, things like website presence, review count, ad-pixel presence, hiring activity — and reduces them to a single number that sorts the call sheet. The agency owner's Friday-afternoon question becomes 'who do I call first?' instead of 'who am I going to skip?'

Two scoring philosophies exist. Heuristic scoring uses fixed weights derived from human judgment ('25% review count, 20% no website, 15% page speed...') and works well at the start of an agency's data journey when there are no closed-won outcomes to train against. Statistical scoring uses logistic regression or a tree model trained on actual close-rate data and outperforms heuristic once the agency has 50+ closed-won outcomes per vertical.

Most local-services agencies live on heuristic scoring for their first two years, then migrate as they accumulate enough outcome data to train against. The discipline matters more than the algorithm — applying the same score to every prospect is what makes the call sheet repeatable.

Outcome Tracking

Logging the eventual disposition of each prospect (closed-won, closed-lost, ghosted) back against the prospect data — the prerequisite for any kind of statistical scoring or attribution analysis.

Outcome tracking sounds trivial until you try to do it across the typical agency stack: prospects come from a list tool, outreach happens in another tool, calls happen on a dialer or by hand, replies come into one of three inboxes, deals close in the CRM. By the time a prospect becomes a customer, the original prospect-data context (score, vertical, signals) is usually lost.

The fix is a stable identifier that flows through every system. Most well-built prospecting tools attach a unique ID to each prospect record on export, then rely on webhook or CSV-uploaded outcomes from the CRM to close the loop. The agency configures the CRM to fire a webhook on each stage transition with the prospect-id custom field.

Without outcome tracking, two things break: the agency can't validate whether the score actually correlates with close-rate (so they can't trust it), and the agency can't train a statistical model later (so they're stuck on heuristics forever).

Line Type Detection

Classifying a phone number as mobile, landline, or VoIP by querying carrier databases — critical for distinguishing dialable real numbers from defunct VoIP forwards.

Line type detection (also called number lookup or carrier lookup) takes a US phone number and returns the carrier's classification: mobile, landline, fixed-VoIP, or non-fixed-VoIP. The data comes from real-time queries against telecom carrier records, routed through specialist carrier-lookup APIs that license access to the underlying number-portability databases.

Why it matters for prospecting: roughly 30% of any local-business phone list is unreachable when dialed cold. Many are non-fixed-VoIP numbers from defunct registrars, abandoned business lines, or fax-only landlines. Without line-type filtering, an agency's outbound caller spends the first hour of every dial session on dead numbers.

For SMS-based outreach, line type is even more critical — sending a marketing SMS to a landline incurs telecom charges and surfaces no engagement. Filtering to mobile-only before the first send saves both reach and money.

Mobile vs Landline vs VoIP

The three primary phone-line classifications — mobile (cellular), landline (POTS or modern equivalent), VoIP (internet-routed) — each with different reachability and cost implications for outbound agency calls.

Mobile numbers are the gold standard for agency outbound: they reach an actual human, are SMS-capable, and the owner is the most likely person to respond. Landline numbers reach the business front desk during hours but answer rates drop ~60% after 5pm. VoIP numbers split into two flavors: fixed-VoIP (provisioned to a physical address, behaves like a landline) and non-fixed-VoIP (Google Voice, Skype, defunct registrars).

The non-fixed-VoIP bucket is the silent killer of cold-call lists. Many older business listings carry non-fixed-VoIP numbers that were valid five years ago but now ring into voicemail nobody checks. Filtering them out before dial reduces wasted dials by 20-30% on a typical local-business list.

For agencies: pull the mobile bucket first for SMS-or-call sequences, the fixed-VoIP and landline bucket second for daytime call sequences, and skip non-fixed-VoIP entirely on cold lists older than 12 months.

DNC Scrubbing

Matching phone numbers against do-not-call registries and known-litigator lists before outbound dialing, and suppressing the ones that come back flagged.

DNC scrubbing matches a phone number against the do-not-call lists that apply to it — the federal National Do Not Call Registry, the registry for the number's own state, the industry opt-out list, and commercially maintained known-litigator lists — and suppresses the matches before anyone dials or texts. It is a matching job rather than a validation job: a clean result says the number is not listed, not that the number is reachable or worth calling.

The check is perishable, because people add their numbers to the registries continuously — so a list is re-checked on a schedule rather than cleaned once. The step-by-step method, the re-check cadence, and the difference between holding the registry data yourself and paying a provider per check are covered in full on the guide below.

How to scrub a call list against the DNC registry — the full guide →

Litigator Trap

A phone number registered to a known TCPA plaintiff who files lawsuits over unwanted calls — one dial is enough to start a claim.

Litigator traps are real phone numbers belonging to professional TCPA plaintiffs. These individuals (sometimes called serial filers) have filed dozens of TCPA suits each, often by deliberately publishing their numbers on directories that cold-callers scrape, then suing on every dial that reaches them.

The traps are concentrated by state — California, Florida, and Texas have the highest litigator counts. Some are listed under business-sounding names that look like normal prospects on a Google Maps pull. Without a litigator-database scrub, an agency's first contact with a trap is a demand letter.

Compliance providers maintain litigator lists and flag any match alongside the federal/state DNC scrub. Adding the layer costs nothing beyond the scrub you are already running; skipping it means the cheapest possible outcome is settling to make a filing go away, and the price of that is set by the other side rather than by you.

TCPA Compliance

Adhering to the Telephone Consumer Protection Act — the federal law governing outbound calls and texts in the US, enforcing consent rules, calling-time windows, and DNC respect.

The TCPA (Telephone Consumer Protection Act, 1991, amended through 2024) is the federal law that governs phone-based outreach in the US. It enforces three big-impact rules for cold outbound: consent before auto-dialing or sending marketing texts, calling only between 8am-9pm in the recipient's local time zone, and honoring federal + state DNC registrations.

The exposure is real, and the thing to understand about it is the unit: it is assessed per contact, not per campaign. A single mishandled list is not one mistake, it is one mistake per number on it, and conduct found to be deliberate is treated more seriously than an accident. State law layers on top of the federal floor and several states are stricter. We do not publish penalty figures here — the amounts turn on which provision was breached, on the state, and on the year, and a number stated confidently and wrongly is worse than no number at all. Price the risk with a telecom-law attorney before you scale outbound, not from a glossary.

How the rules land also depends on the line type and on how the call is placed, and those boundaries are contested rather than settled. Assuming a business landline is automatically fair game is the specific assumption that gets agencies into trouble. What is not contested is the operating posture: human-dialed rather than auto-dialed, DNC-scrubbed before the first dial, time-zone-aware, with a one-step opt-out on every text. That posture is cheap. A wrong reading of an exemption is not, and it is a question for counsel rather than for a vendor.

Cold Email Warmup

The 4-6 week process of gradually ramping email-sending volume on a fresh inbox to build sender reputation with the major mailbox providers — a prerequisite for cold campaigns.

A brand-new inbox (martin@youragency.com) sending 100 cold emails on day 1 will land in spam at Gmail, Outlook, and Yahoo within hours. The mailbox providers grade sender reputation on a track record of legitimate sends, recipient engagement, and complaint rate. A fresh sender has none of those signals — and gets the default treatment, which is the spam folder.

Warmup mimics legitimate engagement. The inbox sends and receives email with other warmup-network inboxes for 4-6 weeks, gradually scaling daily volume from 5-10 sends to 30-50 sends. The exchanges look like normal conversation to mailbox providers; reputation accrues; by week 6 the inbox is ready for production cold campaigns at 30-50 sends per day.

Skipping warmup is the most common mistake in cold-email programs. The cost is not just deliverability on day one — it's permanent reputation damage that takes 60-90 days to repair, plus burn-out of the email domain itself.

Reply Attribution

Linking an inbound email reply back to the originating cold-outbound campaign and prospect — required for measuring per-prospect close-rate and feeding outcome tracking.

Reply attribution is technically simple but operationally fragile. The standard pattern: every cold send carries a unique BCC token (reply+abc123@yourdomain.com) that, when the recipient hits Reply, gets included in the to-line. A poller monitors the bcc inbox, extracts the token, and links the reply back to the prospect record.

Without reply attribution, an agency can measure send volume and aggregate reply rate, but cannot answer the harder question: which prospects replied? Without that, no feedback loop into scoring, no per-vertical reply benchmarks, no outcome tracking past the campaign level.

Practical setup: pick a domain dedicated to reply-tracking (often a sub-domain like reply.yourdomain.com), generate one token per business in the export, and run a poller against the IMAP or Gmail inbox to ingest replies. Mature tools do this automatically; rolling it manually requires ~2 days of one-time setup plus light ongoing maintenance.

CRM Import CSV Format

The expected column layout when pushing a prospect list from a discovery tool into a CRM — varies by destination (GoHighLevel, HubSpot, Salesforce) and is the source of most failed imports.

Every CRM expects its own column layout on a CSV import. GoHighLevel wants firstName, lastName, phone, email, address1, city, state, postalCode, customField.* for any non-standard field. HubSpot wants Email, First Name, Last Name, Phone Number, Company, Lifecycle Stage, plus per-property column headers for custom properties. Salesforce wants flat field names with strict character limits.

Source-of-truth files in well-built prospecting tools include format-specific export buttons. The user picks GoHighLevel CSV instead of Generic CSV, the file is pre-mapped, the import works first try. Without these presets, every export starts with 30 minutes of column-renaming in a spreadsheet.

Custom fields are where things break. Each CRM has limits on custom-property creation; some require pre-creating the property before the import; others auto-create but with the wrong type. Practical fix: create the custom properties in the CRM ahead of the first import (vulnerability_score, line_type, dnc_status, vertical) so the columns map cleanly.

Niche Saturation

The condition where a local market has more agencies competing for a vertical than the underlying business population supports — agency outbound becomes a price war and the win-rate craters.

Niche saturation is the silent killer of agency growth. A market with 200 plumbers and 40 marketing agencies competing for them is fundamentally different from a market with 200 plumbers and 5 agencies. In the saturated case, every prospect has been pitched 3-5 times this year; the price has been bid down to commodity; and the conversation starts with 'we already have an agency.'

Visible signs of saturation: prospects respond to outbound with 'we tried that already, didn't work,' the same business has multiple competing agency pixels on their site, retainer rates in the market trend below $1,500/mo, and the local Reddit / Facebook / LinkedIn community has heated arguments about agency-marketing ethics.

The strategic move on saturation isn't to compete harder — it's to find an adjacent vertical or adjacent market with the same buyer profile but less coverage. A plumbing agency in saturated Austin could pivot to roofing-or-electrical in Austin, or plumbing in San Antonio, both before fighting another year on price.