The Database Accountability System

# Real Estate CRM Database Audit Report
## Purpose
You are analyzing a real estate CRM export (CSV or spreadsheet file) located in this
folder (or uploaded to this session). Your job is to produce a professional Word document
(.docx) called "Database Audit Report" saved to this same location. The report must
include embedded charts and data visualizations generated as PNG images and placed
directly inside the document.
The audience is a real estate team leader who will use this report in 1-on-1 meetings and
team accountability sessions. Write in direct, plain language. Name names. Use specific
numbers. Do not hedge or generalize.
---
## Step 1: Data Discovery and Column Mapping
Before running any analysis, you must first scan the CSV file and map its columns to the
data fields below. Different CRMs use different column names (e.g., one CRM uses "Stage"
while others may use "Status" or "Pipeline Stage"). Read the header row, sample a few
data rows, and identify which columns best match each required field.
Required fields to map:
1. CONTACT NAME - The person's name. Look for columns like: Name, Full Name, First Name +
Last Name, Contact Name, Person.
2. DATE CREATED - When the lead entered the system. Look for: Date Added, Created Date,
Date Created, Registration Date, Created At, Add Date.
3. PIPELINE STAGE - The lead's current stage or status. Look for: Stage, Status, Pipeline
Stage, Lead Status, Contact Status, Deal Stage. Note the exact stage names used in this
CRM, as they vary (e.g., "Showing Homes" vs. "Active Buyer" vs. "Touring").
4. ASSIGNED AGENT - Who owns the lead. Look for: Assigned To, Agent, Assigned Agent,
Owner, Lead Owner, Assigned User. If a value appears to be a team name or company name
rather than an individual person, treat it as an "unassigned pool."
5. LEAD SOURCE - Where the lead came from. Look for: Lead Source, Source, Origin, Lead
Type, Marketing Source, Campaign.
6. TAGS - Any labels or categories. Look for: Tags, Labels, Categories, Groups.
7. LAST ACTIVITY DATE - The most recent interaction. This may be a single column (Last
Activity, Last Contact, Last Touch, Last Communication Date) OR it may need to be derived
from multiple activity columns (Notes, Calls, Texts, Emails, Messages, Activity Log). If
activity is stored as text entries with embedded dates (e.g., "Agent texted Contact on
2/26/2026"), parse the most recent date from each activity column and use the latest one.
Also check for columns like Last Assigned, Last Updated, or Last Modified. Use whichever
date is most recent. If no activity date can be determined, fall back to the Date
Created.
8. ACTIVE PIPELINE STAGES - After scanning the Stage/Status column, identify which stages
represent leads who are actively working toward a transaction. Common examples: "Showing
Homes," "Submitting Offers," "Active Buyer," "Touring," "Under Contract," "Listing
Agreement," "Active Listing." These are the stages used for the Lost Profits analysis. If
you are unsure which stages qualify, use any stage that implies the lead is past initial
contact and actively engaged in buying or selling.
9. EARLY/DORMANT STAGES - Identify which stages represent leads still in the earliest
part of the pipeline. Common examples: "Lead," "New," "Attempted Contact," "Uncontacted,"
"New Lead," "Inquiry." These are the stages used for the Sleeping Database analysis.
After mapping, silently proceed to analysis. Do not ask me to confirm the mapping. Use
your best judgment.
---
## Step 2: Analysis and Report Structure
The report must contain exactly these sections in this order:
### Title Page Area
- Report title: "DATABASE AUDIT REPORT"
- Subtitle: "[Team/Brokerage Name] - CRM Database Analysis" (infer the team name from the
data, such as the most common value in the Assigned Agent pool, the filename, or any
team-level identifier you find)
- Report date (use today's date)
### Executive Summary
One paragraph, 4-6 sentences. State the total database size, the percentage inactive for
90+ days, the number of active pipeline leads going cold, the number of sleeping/
forgotten leads, and the single biggest operational issue you found. This paragraph
should be blunt and specific enough that the team leader can read it out loud and the
team immediately understands the problem.
### Section 1: Database Health Overview
Contents:
- Total contact count (bold)
- Table: Stage breakdown with columns: Stage | Count | % of Total. Sort by count
descending.
- Chart: Horizontal bar chart showing contacts by pipeline stage. Color: light blue
(#6FB3E0).
- Subheading: "Inactivity Summary"
- Table: Three rows showing contacts with zero activity in 30+ days, 60+ days, and 90+
days. Columns: Inactivity Window | Contacts | % of Total. Highlight the 90+ row in red.
- Chart: Bar chart showing the three inactivity buckets (30/60/90 days) with counts and
percentages labeled on each bar. Colors: orange for 30+, red for 60+, dark red for 90+.
### Section 2: Lost Profits Flag
Section heading should be styled in red to signal urgency.
Definition: Leads in active pipeline stages (identified in Step 1, field #8) with no
activity in 30+ days. These are leads who were close to a transaction and are now going
cold.
Contents:
- Introductory paragraph explaining which stages you identified as active pipeline
stages, why these leads matter, and the total count of cold leads.
- Chart: Bar chart showing cold lead count per agent (exclude the unassigned pool from
this chart). Color: red. Label each bar with the count.
- Note (bold, red text) stating how many of these leads are sitting in the unassigned
pool and need immediate reassignment.
- For each agent who has cold leads: a subheading with the agent's name, followed by a
table with columns: Lead Name | Stage | Days Inactive | Lead Source. Sort by days
inactive ascending. Color-code the Days Inactive values: 30-59 days = default color,
60-89 days = orange, 90+ days = red and bold.
### Section 3: Agent Accountability Breakdown
Contents:
- Introductory paragraph noting how many leads are in the unassigned pool (excluded from
individual agent counts).
- Summary table with columns: Agent | Total Leads | Inactive 30+ Days (with %) | Inactive
90+ Days (with %) | Most Recently Active Lead (name and days) | Most Dormant Lead (name
and days). Sort by total leads descending. Color the inactive percentage cells red if
they exceed 80%.
- Chart: Horizontal stacked bar chart showing active leads (light blue, #6FB3E0) vs.
dormant leads 90+ days (red, #CC0000) for each agent. Label the right side of each bar
with the dormant percentage. Exclude agents with fewer than 50 leads for chart
readability.
### Section 4: Sleeping Database Summary
Definition: Contacts created more than 6 months ago, still in early/dormant pipeline
stages (identified in Step 1, field #9), with no activity in 90+ days. These are
forgotten leads.
Contents:
- Introductory paragraph stating the total count of sleeping leads and what percentage of
the full database this represents.
- Subheading: "By Stage" with a two-column table: Stage | Count
- Subheading: "By Agent Assignment" with a two-column table: Agent/Pool | Sleeping Leads.
Include all agents and the pool. Sort by count descending.
- Subheading: "By Lead Source (Top 10)" with a two-column table: Lead Source | Count
- Chart: Horizontal bar chart of sleeping leads by lead source (top 8). Color: deep navy
(#0B1F3A).
### Section 5: Action Recommendations for Your Next Team Meeting
Contents:
- Brief intro sentence: "The following five items are written in plain language so you
can read them directly to your team or use them as agenda points in your next 1-on-1s."
- Exactly 5 recommendations, each with a bold numbered subheading and a paragraph of 3-5
sentences. Each recommendation must reference specific numbers from the data. These
should address: (1) reassigning pool/unassigned leads, (2) triaging the lost profits
list, (3) setting a team standard for scheduling next steps on every lead, (4) launching
a re-engagement campaign for the sleeping database, and (5) holding agents accountable to
their dormant-to-active ratio using the data in this report.
---
## Document Formatting Rules
### Brand Identity
This report uses the brand palette defined below. All colors, typography, and styling
must follow this guide consistently.
### Typography and Colors
- Font: Arial throughout the entire document
- Body text: 11pt, color #333333 (Dark Gray)
- Section headings (H1): 16pt, bold, color #1D6FB8 (Blue - primary brand color)
- Subheadings (H2): 13pt, bold, color #1D6FB8 (Blue)
- Report title: 20pt, bold, color #0B1F3A (Deep Navy)
- Report subtitle: 12pt, color #3D4E5C (Slate Blue)
- Page header: right-aligned, italic, 9pt, color #999999: "[Team Name] - Database Audit
Report"
- Page footer: centered page numbers, 9pt, color #999999
- The Lost Profits Flag section heading (Section 2) should remain red (#CC0000) for
urgency
### Page Layout
- US Letter size (8.5 x 11 inches)
- 1-inch margins on all sides
- Page breaks before Section 2, Section 3, Section 4, and Section 5
### Tables
- Header row: blue background (#1D6FB8), white bold text
- Alternating row shading: light gray (#F5F5F5) on even rows
- Light gray borders (#EEEEEE)
- Cell padding for readability (internal margins of 60-100 DXA)
- Right-align all numeric columns
- The 90+ inactivity row should use a light red background (#FFF3F3) with red text
(#CC0000)
### Charts
- Generate all charts as PNG images using matplotlib (or equivalent) at 180 DPI
- Embed each chart directly in the Word document as an inline image
- Use a clean, professional style: light background (#F5F5F5), subtle gridlines (#EEEEEE
at 30% opacity), no chartjunk
- Primary color palette: blue (#1D6FB8), light blue (#6FB3E0), deep blue (#0B4F8A), deep
navy (#0B1F3A)
- Alert/urgency palette: red (#CC0000), orange (#CC6600), dark red (#880000)
- All chart titles: 14pt bold, blue (#1D6FB8), with padding above
- Data value labels on bars: deep navy (#0B1F3A), bold
- Label data values directly on bars where possible
- Center all chart images in the document
### Section Dividers
- Use a thin blue horizontal rule (#1D6FB8) between major sections
---
## Output Requirements
- Save the final file as "Database Audit Report.docx" in this folder (or this session)
- The document must open cleanly in Microsoft Word, Google Docs, and Apple Pages
- All charts must be embedded (not linked) so the document is fully self-contained
- Do not ask me any questions before starting. Scan the CSV, map the columns, run the
analysis, and produce the report.

You are my real estate operations analyst. I just uploaded a CSV. It is a full lead
export from my CRM for one of my agents. I have not cleaned or filtered it. Read every
row. Cross-reference activity, history, visits, tags, status, and notes on your own.
Before analyzing, scan the header row and identify what data this export actually
contains. CRMs differ - some track lead scores, dialer activity, login history, and
property views; others track only stages, sources, and communication logs. Build the
report from the fields that exist in this file. If a useful metric is not in the export,
leave it out rather than guessing. If a field exists but is empty for a lead, mark it as
Unknown rather than skipping the lead.
Produce an Agent Roadblock Report. Use plain English, short sentences, hyphens instead of
em-dashes, and a confident, no-fluff tone. Pull every number, name, date, and address
straight from the CSV - do not invent anything.
The report should be organized into three parts. You can adjust formatting (tables,
bullets, paragraphs) to whatever reads cleanest, but make sure all of the following
information is captured and clearly presented.
PART 1 - PERFORMANCE SCORECARD
A snapshot of the agent's activity inside the CRM. Include the metrics that are useful
for spotting dormancy, missed follow-up, and overall pipeline health, drawn from whatever
this export supports. At minimum cover: total database size, dormancy rate (leads with no
activity in 90+ days), cold or lost active leads, new leads in the last several weeks,
leads not yet contacted or viewed, and outbound call/text/email counts in the last
several weeks. Where the export supports them, also cover: year-to-date closed
transactions and volume, recent session or login activity, time to first email, time to
first call, dialer activity, lead scores, and tasks sitting with no action. Flag each
metric as OK, Warning, or Critical based on whether it is healthy, slipping, or genuinely
broken. Use your judgment on thresholds based on what looks normal vs. neglected in the
data.
PART 2 - HOTTEST LEADS RIGHT NOW
The top handful of leads (5 is a good default) most likely to convert in the near term.
For each one, capture name, status, days since last touch, and source. Pick based on a
combination of recent engagement signals (visits, properties viewed, replies, lead score,
recent activity - whatever this export tracks) and viable status. Rank them in priority
order.
PART 3 - PERSONALIZED FOLLOW-UP SCRIPTS
For each of the leads listed in Part 2, give the agent everything they need to make the
call. Capture:
- Lead name, type (buyer, seller, nurture), and source.
- Phone, email, and location.
- Price band, visit count, and properties viewed (if tracked).
- Key Intel: a short paragraph pulling specific facts from the CSV - what they clicked,
what the agent has already done, dates of last touches, referral source, lead score,
anything that explains why this lead is hot or stuck. Use real dates and real properties
from the data.
- A recommended script or approach the agent can read off their phone. Reference the last
specific interaction. Sound human, not corporate. If the lead has been called multiple
times with no callback, pivot to text or email and write the actual message to send.
ASK ME FIRST
Before producing the report, ask:
1. Agent's first and last name.
2. Their role (listing, buyer, dual).
3. Any context I want factored in (new agent, slump, just back from leave, personal
stuff, etc.).
Then read the CSV and produce the report.

You are my real estate CRM consultant and copywriter. I am going to share information
about my business, then you will help me (1) audit my current CRM and (2) build three
human-sounding, 120-day action plans for Buyer leads, Expired listing leads, and Home
Value / CMA leads.
Before you do anything, ask me the following questions one at a time and wait for my
answer before asking the next:
1. What is your name, brokerage or team name, and market area (city, state)?
2. Which CRM do you use? (Follow Up Boss, kvCORE, BoomTown, Chime, Real Geeks, Lofty,
Sierra Interactive, HubSpot, other)
3. How many active leads are in your CRM right now? Roughly what percent have a current
status tag?
4. What are your three biggest lead sources (Zillow, Google PPC, Facebook, referral, open
house, etc.)?
5. How do you currently follow up with new leads on Day 1? Walk me through the exact
sequence.
6. How long do your current drips run? Do they feel robotic to you?
7. Who assigns leads to agents, and how? (Auto round-robin, manual, shared queue, direct)
8. What is your personal communication style? Formal, casual, funny, no-nonsense?
9. Do you currently send any physical mail? (Mailers, handwritten notes, CMAs)
10. What is the one thing you wish your follow-up did better?
After I answer all 10 questions, do the following in this exact order:
Part A - CRM Audit Report
- Give me a 1-page summary of what is working and what is broken in my current setup.
- Score me 0-10 on Data Hygiene, Automation Discipline, Agent Accountability, Cadence
Quality, and Tone. Be honest.
- List the top 3 things I should fix before building new plans.
Part B - The 3-Plan Build
For each of the three plans (Buyer, Expired, Home Value / CMA):
- Confirm the 120-day structure with activity count and cadence.
- Write the full copy for every Text, Email, and Voicemail script. Match my tone from
question 8. Reference my market area from question 1. Use hyphens only, never em-dashes.
- For each activity, include three things in the Notes field: Suggestion, Example, Pro
Tip.
- Flag which 2 activities should have auto-send ON (Day 1 intro text and email only, both
containing "A real agent will be following up shortly").
- Everything else is manual send.
Part C - Implementation Checklist
- Step-by-step build instructions tailored to my specific CRM from question 2.
- A weekly maintenance rhythm (15 minutes per week minimum).
- A 30-day review plan so I can measure if the plans are working.
Rules you must follow:
- No em-dashes. Hyphens only.
- No robot-speak. Write like a real agent talking to a real human.
- Every text must be under 160 characters unless specifically noted.
- Every email must be scannable: short paragraphs, one clear ask.
- Every voicemail script must be under 25 seconds when read aloud.
- If I give you a market area, reference local specifics where possible (neighborhoods,
school districts, price points) but do not invent fake data. Mark placeholders like
[local neighborhood name] when needed.
- Confirm with me before you start Part B. I want to approve the audit first.
Ready? Start with question 1.