Most businesses do not have a lead problem.

They have a data problem.

Before you can call, email, advertise, follow up, build a CRM, or create a sales pipeline, you need clean contact data.

This free training shows you how to find business data, clean it, structure it, and use AI carefully to build a practical lead generation system.

You will learn how to work with Google Maps, spreadsheets, ChatGPT, Gemini, CRM ready columns, names, addresses, emails, mobiles, websites, and real business follow up notes.

What this page teaches

• How to find business contact data

• How to use Google Maps for lead generation

• How to clean messy spreadsheet data

• How to use AI without trusting it blindly

• How to build CRM ready rows

• How to prepare data for calling, email, advertising, and follow up

Start learning below, or buy a ready-built starter data pack if you want the work done for you.

Want the business data built for you instead? Start with a $299 BizConnector data pack on Trade Me.

Free Training Step 1

How to think about business data

Most businesses think they need more leads, but first they need a clean list of the right people or businesses to contact.

Good data helps you decide who to call, who to email, who to advertise to, and who to follow up.

Bad data wastes time, causes missed follow ups, and makes sales teams blame the calls when the real problem is the list.

Free Training Step 2

Start with the right search

Before you build a database, decide exactly who you want to find.

You can build a very large database, but start with one clear industry and one clear region first.

Using Google Maps to search by industry is a great starting point.

Example searches: use clear business search terms

• Real estate agents in Nelson

• Builders in Palmerston North

• Accountants in Tauranga

• Cafes in Christchurch

• Property managers in Auckland

• Electricians in Hamilton

Now open a Google Sheet. Keep each sheet under about 300,000 rows, then place these headers into the header row.

• Registry Name, if sole trader, use their mobile number with no spaces and no leading 0

• NZBN Number, if sole trader, write Sole Trader or Private

• Business Legal Name

• Trading Display Name

• Category

• Full Address, example 40 Lethbridge Street, Feilding, Manawatū, 4702

• Post Code

• Suburb or Town, example Feilding

• Region and Town, Top 20 cities only, example Manawatū

• Phone Number, example 09 268 0282

• Mobile Number, example 022 637 8387

• Email, example dane@funnytelemarketer.nz

• Social Sites, example Facebook or LinkedIn

• Website, example www.funnytelemarketer.nz

• Company Description

• Name Last, example Wells

• Name First, example Dane

• Position, example Director

• Status, example client, follow up, BDM, appointment, 26 to 35 for yearly follow up, or 26.bdm

• Appointment or follow up information, add appointment date, time, BDM call back, or follow up month from 1 to 12

• Notes, start with the date and write newest notes first, example 26.05.10 Dane spoke with the owner and they asked for more information

Free Training Step 3

Use ChatGPT and Gemini together

I think of ChatGPT like C-3PO. It is good at structure, systems, wording, instructions, and building the SOP.

I use Gemini more like the web search and data checking tool. Gemini is useful for searching, checking public information, and helping clean the data into the format you want.

The simple process is:

1. Use ChatGPT to write your SOP.

2. Put that SOP into Gemini as a Gem.

3. Use Gemini to return the data in the exact columns you asked for.

4. Ask Gemini to return the final output in a raw TSV block grid.

5. Copy the TSV block and paste it straight into your spreadsheet.

Example of what your SOP instruction for Gemini could look like

Important: this is only an example SOP. You can start simple and improve it over time. Do not try to make the first version perfect. The goal is to build a working system first, then patch it when you see mistakes.

Below is an example of a more advanced SOP patch.

You do not need to understand every line at the start. This shows how specific the instructions can become once your data system grows.

Start simple first. Then patch the SOP when Gemini makes repeat mistakes.

26.05.07 patch

THE PATCH: Upgraded Deep Search & Verification Instructions

1. Expanded "Deep Search" Methodology (Replace Original Step 1)

When provided with a raw business data line, you must perform an exhaustive, multi-layered search:

Primary Web & Social Scrape: Search the official website ("Meet the Team," "About Us," "Contact"). You must also scrape the company's official Facebook/LinkedIn pages for staff mentions, "employee of the month" posts, or administrative contacts listed in social bios.

Group & Entity Cross-Reference: Look up the business on the New Zealand Companies Office. Do not stop at the primary entity. Search the Directors' names to find affiliated businesses, sister companies, or parent groups operating from the same address or region. Include any overlapping Group Directors or Partners.

Admin & Front Office Sweep: Deliberately search for non-director operational roles. Search business directories (Yellow Pages, Localist, industry associations) specifically for Office Managers, Admin Leads, and Service Managers.

Row Duplication: Duplicate the company data line so every individual (Director, Owner, Group Partner, Manager, Admin, and Key Staff) gets their own dedicated row.

2. Adjusted Verification Rules for Auxiliary Staff (Add to Step 2)

Group Affiliation Verification: If a staff member (e.g., a Group Director) is not legally registered to the exact NZBN but is a verified partner of the overarching regional brand, include them and note the group affiliation in Column 26.

Admin Verification: If front-office staff (e.g., Office Manager) lack a direct, published email, you may leave the email blank () but you MUST still extract their name and role based on verified social or directory mentions.

Custom GEM System Prompt: Deep-Wash & Verification Specialist

Core Persona & Objective

You are a high-level Data Extraction and Verification Specialist. Your primary task is to take raw company data strings, perform a "deep search" across the web, and output a highly accurate, deep-washed TSV (Tab-Separated Values) grid. You prioritize data accuracy, strict verification, and exhaustive staff extraction over speed. You never fabricate data.

1. The "Deep Search" Methodology

When provided with a raw business data line, you must perform the following steps:

Website Scrape: Search the company's official website, specifically targeting "Meet the Team," "About Us," or "Contact" pages to identify all visible staff.

Companies Office Cross-Reference: Look up the business on the New Zealand Companies Office (or OpenCorporates) to identify all officially registered Directors and significant Shareholders.

Row Duplication: For every unique person identified (Director, Owner, Manager, or Key Staff), duplicate the company data line so every individual has their own dedicated row in the final TSV output.

2. Strict Verification & Anti-Fabrication Rules

Data integrity is paramount. You must adhere to the following rules when populating contact information:

No Guesswork: You must never fabricate, engineer, or guess an email address or mobile number. If you cannot find a verified link, leave the field blank using ().

Email Verification: * Only extract emails explicitly listed on the company website, official industry registries (e.g., SGNZ, Master Builders), or verified LinkedIn profiles.

Domain Check: Always verify the active operational domain (e.g., catching if a company uses .co.nz instead of .com).

Mobile Verification: Ensure the mobile number connects to the actual person listed. Cross-reference numbers to eliminate generic redirect numbers or false positives from other businesses.

The Verification Note: You must explicitly state how the person and their contact data were verified in Column 26 (e.g., "Verified: John Doe is Managing Director. Email confirmed via official site domain.").

3. The 28-Column Formatting Rules

You must output the final data strictly in a raw TSV block. Do not use markdown tables. Each row must contain exactly 28 columns separated by tabs. Use () for any blank or unverified fields.

Column Mapping:

ID: Internal ID (Leave as provided or ())

NZBN: NZ Business Number

Blank: ()

Blank: ()

Legal Name: Official registered company name (e.g., LEISURECOM (NZ) LIMITED)

Trading Name: Brand name (e.g., Leisurecom Homes)

Industry: Category (e.g., Trades / Builder)

City: Main operating city

Postcode: 4-digit code

Address: Physical operating address

Email: VERIFIED direct email only (or ())

Phone: Landline/Office number

Blank: ()

Mobile: VERIFIED direct mobile only (or ())

Toll-Free/Other: e.g., 0800 numbers (or ())

Social URL: LinkedIn/Facebook link (or ())

Website: Official verified active domain

Year Founded: Year of incorporation

Description: Brief summary of operations

Region: e.g., Waikato, Cambridge

Last Name: Staff surname

First Name: Staff given name

Job Title: Official role (e.g., Director / Finance)

Status: e.g., Registered

Blank: ()

Verification Notes: Explicit detail of how data was verified. MUST start with "Verified: "

Employee Count: e.g., 11 to 20

Revenue Range: e.g., $5M - $10M

4. Output Formatting

Provide a brief summary of what was found and verified.

Present the data inside a tsv ...  code block.

Do not ask follow-up questions at the end of the data block. Simply provide the washed data.

How to load this into Gemini as a Gem

Open Gemini.

Go to Gems.

Create a new Gem.

Name it something simple, for example:

Business Data Cleaner

Paste your SOP into the Gem instructions or description box.

Save the Gem.

Now, when you paste scraped business data into that Gem, it should return the information in the columns you asked for.

How to collect the first data

Open Google Chrome.

Go to the Chrome Web Store.

Search for:

Instant Data Scraper

The logo looks like a small Pokémon ball.

Add it as a Chrome extension.

Then open Google Maps.

Search for the type of businesses you want, for example:

builders in Palmerston North

Open Instant Data Scraper.

Scrape the business list or the visible tiles it finds.

Move the map to a new area and scrape again.

You can also click the next pin or move through the list to collect more results.

Instant Data Scraper can hold up to about 1,000 data lines.

Important: scraped data must still be checked. Do not trust every phone number, website, email, or business name until it has been cleaned and verified.  (tell gemini on regular basis you must check the website for verifed information only please ) when it learns to do this then you can say check social like face book and linkidlin for emails and mobile numbers please 

Example: Instant Data Scraper collecting business data from Google Maps

In this example, I searched for builders on Google Maps, opened Instant Data Scraper, and let it collect the visible business results.

The scraper is pulling basic data such as business name, category, address, phone number, website, and Google listing links.

This is the raw data stage. It will not be perfect yet.

The next step is to copy the scraped data, paste it into the Gemini Gem, and ask Gemini to clean it into the correct spreadsheet columns as a raw TSV block grid.

Free Training Step 4

Copy the raw scrape into Gemini

Once Instant Data Scraper has collected the business list, copy the data and paste it into your Gemini Gem.

Ask Gemini to clean the data into your spreadsheet format.

Use this prompt:

Please clean this scraped business data into my SOP columns.

Return the result only as a raw TSV block grid.

Use tabs between each column.

Do not use a markdown table.

Do not guess missing information.

If something is missing, write Unknown.

Keep the column order exactly the same as my SOP.

Here is the scraped data:

Once Gemini gives you the raw TSV block grid, copy it and paste it into your spreadsheet. The data should line up with your columns.

Check the rows before you keep going. Look for missing tabs, wrong columns, guessed information, spelling mistakes, and anything that does not look right.

For more specialised data work, you can give Gemini a tighter instruction. For example, you can say you only want businesses in one region, businesses in one category, businesses with a certain number of staff, businesses with directors listed, or businesses that match a particular service.

If you want a deeper business search, create a second SOP or a second Gemini Gem. Copy the cleaned data from the first stage into the new Gem and add the special instructions.

Example:

I need you to do a deeper search on each business. I want director names, staff names where public, verified phone numbers, emails, websites, NZBN numbers, addresses, and useful notes.

Breaking the process into a few smaller Gems usually works better. Gemini can struggle when one Gem has a very large and detailed SOP.