BirdsEye ROI, Sales Enablement

Angle authority is marketing/positioning-why-i-built-it.md. Read it before this document.

Source of truth for every number and capability claim is marketing/FACTS.md. If a rep says a number, it is on that sheet. If it is not on that sheet, the rep describes the mechanism, not a result. This document does not repeat the fact table, it teaches a rep how to run it in a live conversation without going off script.

The one rule that governs how numbers get used. The fact sheet is for ANSWERING, never for OPENING. A rep never leads a call, an email, or a demo with the spam share, the total call count, or the tracked revenue total. Those are facts about our database, not reasons anybody buys. They come out only when a prospect asks a direct question that one of them answers.

A rep opens on the buyer's own problem:

• Operator: you get about four seconds between jobs, and the number should already be on the screen. What today made. What the last call was worth.
• Agency owner: you are not short on calls, you are short on proof. The Friday report, per service and per area, that you are proud to send without editing.

Section 1: Rep onboarding, first 5 days

Each day ends with a checkable output. A rep does not move to the next day until the prior day's output is verified by a manager.

Day 1: Learn the product cold

• Read positioning-why-i-built-it.md first, then FACTS.md twice. Read the primary line out loud three times: "Know what the last call was worth before you put the phone down."
• Learn the four supporting lines: a thirty thousand foot view of your business on the screen you already check; your calls, with a dollar figure attached; the report your client actually wants to see; four seconds, and you know if today paid.
• The line "most call tracking tells you the phone rang, BirdsEye tells you which ring paid" is RETIRED. It argues with a category instead of talking to a person. It stays in the deck and never comes out of a rep's mouth on a live call.
• Log into the live product at https://birdseyeroi.com and click through all nine pages: Calls, Dashboard, Leads, Categories, Calendar, Numbers, Hawkeye, Texts, Alerts.
• Checkable output: state the primary line and both buyer openers unaided, then recite 8 of the 12 verified numbers from the fact table from memory. The numbers are answering ammunition. A rep who opens a mock call with one fails the day.

Day 2: Learn who this is for and why

• Read Section 2 of this document (the discovery framework) for both buyer types.
• Shadow two live or recorded discovery calls, listen only, no talking.
• Checkable output: write one sentence each on what problem a local service owner has and what problem a lead-gen agency owner has, in the rep's own words, matching the framing in Section 2.

Day 3: Learn to handle friction

• Read Section 3 (objection drills) and Section 4 (overpromise guardrails) once each.
• Run a mock discovery call with a manager playing a local service business owner.
• Checkable output: manager fires 5 objections from the list of 12, rep answers all 5 using the fact sheet backing, zero invented numbers, zero invented claims.

Day 4: Learn where the line is

• Re-read Section 4 until the "never say" list can be recited unaided.
• Run a second mock call, this time as a lead-gen agency owner, including at least one disqualifying signal the rep must catch.
• Checkable output: rep correctly identifies the disqualifier and states out loud why they would walk rather than force the deal.

Day 5: Full unaided demo

• Deliver a complete demo to a manager with no notes: open on the buyer's own problem for the assigned track, show the end of day view for an operator or the per service and per area report for an agency owner, then narrate all nine pages, field three objections drawn at random from Section 3, and run one full discovery call role play scored against the Section 5 rubric.
• The spam-filtering story is available if a manager asks how junk is handled. It is not the centerpiece of the demo and never opens it.
• Checkable output: rubric score of "pass" with zero guardrail violations. A single overpromise violation is an automatic fail regardless of every other score. A rep who fails Day 5 repeats Days 3 through 5, not the whole week.

Section 2: The discovery call framework

Every question exists to qualify one thing. Ask it, then name to yourself what it just told you.

Local service business owner track

Disqualifiers, walk rather than force:

• No real call volume, and none expected. There is nothing to attribute.
• No paid marketing spend anywhere. Attribution has nothing to compare against.
• A true owner-operator who personally answers every call and already knows exactly where it came from. This buyer has no problem yet.
• Already locked into a competing platform under contract and expressing no dissatisfaction.

Lead-gen agency owner track

Disqualifiers, walk rather than force:

• Fewer than a handful of clients and no plan to grow the book. Not enough volume to justify a platform switch.
• Runs display or SEO only, no call-based lead gen. There is no call data to unify.
• Already building a proprietary tracking stack in-house and committed to finishing it.
• Wants call flows sold today as a finished, battle-tested feature. It is not, and Section 4 covers exactly how to have that conversation honestly instead of losing the deal on a lie.

Section 3: Objection handling drills

Each response is backed by a specific fact. Where there is no backing number, the drill says so.

1. "We already have call tracking." Response: most operators are running several providers at once with no single view. This platform unifies 4 (Twilio, SignalWire, CallFire, CallRail) into one call log. Backed by: Verified platform numbers, "Call tracking providers unified."

2. "How do I know your numbers aren't padded with junk calls?" Response: the platform flags spam automatically, 43,700 calls flagged, 42.3% of the full log, and spam is excluded from the revenue math. Backed by: "Calls flagged as spam by the platform."

3. "What ROI will I actually get?" Response: cannot be answered with a promised number, no client result numbers exist to cite. Answer with the mechanism instead: revenue rules apply per call, per service, and per niche, and the ROI panel shows revenue, cost per lead, margin, and ROI once those rules are set for their business. Backed by: no number available, see Section 4 for the honest script.

4. "Does it work with my call provider?" Response: only if it is one of the 4 unified providers, Twilio, SignalWire, CallFire, or CallRail. If it is not, say so plainly rather than imply it will. Backed by: "Call tracking providers unified."

5. "Can it route my calls, IVR, ring groups, that kind of thing?" Response: 13 call flow types are built and wired end to end. None have carried a live call yet. Use the exact guardrail script in Section 4, do not present this as production-proven. Backed by: "Call flow types built and wired," cross-checked against project status (never tested live).

6. "How many calls has this platform actually handled?" Response: 103,000+ calls are in the call log today. Backed by: "Calls in the call log."

7. "Do you get transcripts, or just call logs?" Response: qualifying calls are auto-transcribed with AssemblyAI Universal-2. Backed by: Verified capability list, auto-transcription line.

8. "Can it tell me what the caller actually wanted?" Response: yes, each call carries 13 dedicated AI-extracted fields: caller name, service needed, urgency level, customer type, call outcome, and more. Backed by: "Dedicated AI caller-intelligence fields per call."

9. "What about text messages, not just phone calls?" Response: yes, SMS is analyzed for intent, money signals, and urgency, 7,495 conversations analyzed to date. Backed by: "SMS conversations analyzed."

10. "Can I see revenue broken out by service, not just by lead count?" Response: yes, 79 service monetization records across 18 canonical, normalized service categories, with revenue, cost per lead, and margin at the service level. Backed by: "Service monetization records," "Canonical service categories normalized."

11. "Do you have case studies or references I can call?" Response: no client names, testimonials, or case studies exist to share, and none will be invented. Offer a live walkthrough of the real platform data instead of a fabricated proof point. Backed by: no number or reference available, this is a hard no-bluff line per FACTS.md.

12. "Will an AI answer my calls for me, like a robot receptionist?" Response: AI receptionist is one of the 13 wired flow types. It is built but has never carried a live call. Use the Section 4 script, offer to bring it live together as the first real test. Backed by: "Call flow types built and wired," cross-checked against project status (never tested live, AI receptionist is SignalWire-only).

Objections a rep cannot answer with a number at all: promised ROI, case studies or references, "how many companies like mine use this," win rate against a named competitor, churn rate, total customer count. On every one of these, the rule is the same: say what is real (the mechanism, the platform's own numbers), say plainly that a client-outcome number does not exist yet, and move to a live demo of real data instead of a claim that cannot be checked.

Section 4: The overpromise guardrails

This is the one section every rep must know cold before taking a live call. The fleet only sells what it can ship. A rep who oversells creates a support and trust problem that lands back on the whole team.

Never say, under any circumstance

• Never open a call, email, or demo with the spam share, the total call count, or the tracked revenue total. Those are facts about our database, not reasons anybody buys. They answer a direct question and nothing else.
• Never use the retired line "most call tracking tells you the phone rang, BirdsEye tells you which ring paid" on a live call. Deck only.
• Never claim a specific client's revenue, ROI, or result. No client names, testimonials, quotes, logos, or case studies exist. Do not imply one does.
• Never state a revenue or call-volume number that is not on FACTS.md. Never round up. Never extrapolate ("if you had our system for a year you'd see...").
• Never say the 13 call flow types are "proven," "battle-tested," "in production," or "already handling calls for other clients." They are built and wired, and zero calls have gone through any of them.
• Never say the AI receptionist flow "answers your customers today" as a live, working feature. It exists in code and is wired to SignalWire, and it has never taken a real call.
• Never claim integration with a call provider outside the 4 unified ones (Twilio, SignalWire, CallFire, CallRail).
• Never promise a go-live date for call flows without a manager confirming a test plan exists.

What exists versus what is built but untested, and how to talk about the difference

Fully live and proven on real data (safe to sell as-is):

• Unifying calls across the 4 providers into one log, 103,000+ calls processed
• Spam auto-flagging, 43,700 calls, 42.3% of the log, excluded from revenue math
• Auto-transcription via AssemblyAI Universal-2
• AI sentiment scoring and the 13 caller-intelligence fields per call
• SMS analysis on 7,495 conversations
• Revenue rules and ROI reporting by niche and service, 79 service monetization records across 18 categories

Built and wired, never carried a live call (sell honestly, do not sell as finished):

• All 13 call flow types: simple forward, IVR menu, ring all, hunt group, time-based, overflow, round robin, voicemail, whisper, recording, call queue, API bidding, AI receptionist

The honest script for the second category: "That flow type is fully built, the UI, the database config, and the provider connection are all there and wired to fire on a real call. What has not happened yet is a live call actually going through it. If that matters to you, we bring it live together as the first real test, you watch it happen, we do not hand you something we have not verified with you in the room."

This keeps the conversation open and honest. It never claims something works that has not been proven, and it never kills the deal by hiding the gap either, it turns the gap into a joint next step.

Section 5: Call review rubric

A manager scores each recorded discovery or demo call on the following dimensions, 1 to 4 each. One rule overrides every score: any confirmed overpromise violation (see Section 4) is an automatic fail for the whole call, regardless of every other dimension.

Pass requires an average of 3 or better across the first six dimensions and zero guardrail violations. A guardrail violation fails the call outright even with a perfect score everywhere else.

Section 6: Team ritual

Weekly pipeline review, every Monday, run by the sales lead.

• Every rep's calls from the prior week are logged. The lead pulls one call per rep at random, not the rep's own pick, and scores it against the Section 5 rubric.
• Any call flagged with an overpromise violation is reviewed with the full team in the same meeting, not handled quietly one-on-one. This is not about punishing the rep, it is about keeping the whole team's script current and catching drift before it repeats.
• Pipeline is reviewed stage by stage against the disqualifiers in Section 2. A deal sitting in an early stage with a known disqualifier gets flagged and closed out, not left to rot.

The one metric that cannot be gamed: fact-accuracy audit score. Close rate, call volume, and self-reported objection handling can all be shaped by which calls a rep chooses to report on. The fact-accuracy audit cannot, because the lead picks the call, not the rep, and checks every stated number against FACTS.md directly. A rep's real number is the percentage of audited calls with zero invented or rounded numbers. That score is the one the team tracks over time, because it is the one no rep can quietly optimize by cherry-picking their best call.