CentriWebCRM implementation sticks when you do it in a fixed order: decide your stages first, load your data second, and add automation last. Most failed set-ups do the reverse. They start with a product demo, import everything, switch on a dozen automations and then find nobody trusts the screen. This guide gives the order, what each step should produce, and a first-week plan.
It is written for a business with roughly two to fifteen people who handle enquiries, quotes or bookings and want one place where that work lives. It is not specific to any product, so it works whichever system you pick.
Before you touch the software
Three decisions come before any set-up. Make them on paper in an hour.
- What problem is the CRM for? Write one sentence. “Stop quotes going quiet” is a good one. “Get organised” is not, because you cannot check it. The sentence decides what you build and what you leave out.
- Who owns it? One named person, with the authority to say no to new fields. Not a committee. In a small business this is usually the owner or an office manager.
- Who uses it daily? List the people and what each does in it. A tradesperson on site might only need to see today’s jobs and add a note. An office admin might need everything.
If you are not yet sure you need a CRM at all, read CRM vs spreadsheet: when to switch first.
Who does what in week one
Name the jobs before the week starts, or they fall to whoever is least busy. The owner decides the purpose and signs off the stages. The person who owns the system builds fields, cleans the data and runs the walk-through. Each daily user tests their own screen on their own phone and says what is awkward, on the day, while it can still be changed cheaply. If an outside helper does the build, the owner still approves the stages, because only you know how your customers decide.
Also decide what you will not do yet: reporting dashboards, scoring leads, marketing campaigns and a second pipeline. All of them can wait until the first one works.
Step 1: Define the stages
The pipeline is the spine of the system, so build it before anything else. Each stage should describe something the customer has done or something that has happened. A feeling about the job does not count. “Quote sent” is a fact. “Warm” is a mood.
For most service businesses, six to eight stages is enough. A typical set:
- New enquiry
- Contacted (you have spoken, not just left a message)
- Site visit or consult booked
- Quote or proposal sent
- Following up
- Won
- Lost
Write down what moves a record from one stage to the next, and who moves it. We go through the choices in detail in sales pipeline stages for a service business.
Step 2: Choose the fields, then cut half
Every field is a question someone has to answer, usually on a phone, usually in a hurry. For each field you want, ask: will we make a decision based on this within a month? If not, leave it out. You can add fields later in minutes. Removing them after people have learned to ignore them is harder.
A sensible starting set for a contact:
- Name, mobile, email, suburb
- Source (a fixed list)
- Service wanted (a fixed list)
- Estimated value
- Owner
- Next action and its date
- Lost reason (a fixed list, required when marking Lost)
That is about ten fields. Anything beyond it needs a reason you can state out loud.
Step 3: Clean and import your data
Bring in only what you will use. In most businesses that means current enquiries and open quotes, plus customers you expect to hear from again.
- Export from wherever the data sits now: spreadsheets, email contacts, your phone, a job management tool.
- Combine into one file with one row per person. Merge duplicates by phone number and email.
- Fix the fixed-list columns. Make “google”, “Google” and “Google Ads” one value.
- Delete anyone you would not call. Old, bounced or unsubscribed addresses are a risk and add nothing. Do not load people into a marketing list simply because they are in your address book. Rules on consent for commercial messages differ by country, so check your regulator’s guidance for your situation.
- Import a small test batch of ten. Open each record and compare it with the source.
- Import the rest.
Resist importing years of history to look thorough. A clean three hundred records beat a messy five thousand.
Step 4: Connect enquiry sources
Now point your real enquiries at the system, so new work lands there without anyone typing it. Start with the one or two sources that produce most of your volume, usually the website form and the phone.
- Website forms: send each submission in as a new contact with its source set. Our guide to web forms into a CRM covers the common ways to do it.
- Facebook and Instagram lead forms: see Facebook lead ads into a CRM.
- Email and phone: log them from the CRM’s own tools, or have the person who took the call add a note on the spot.
Test every connection by submitting a real enquiry yourself and watching it arrive.
Step 5: Add automation last, and only a little
Automation is where a CRM earns its keep, and where over-eager set-ups go wrong. Start with three, and no more:
- A task for the owner when a new enquiry arrives, due the same day.
- An instant acknowledgement to the customer, written like a person.
- A reminder when a quote has sat in “Quote sent” for a set number of days.
Add others only after the first three have run for a month and people trust them. For the broader picture of what to automate, see our automation service. Automation built on bad stages or dirty data makes mistakes faster, which is why it comes after steps 1 to 3.
A first-week plan for CRM implementation
This is a realistic week for a small team. It assumes the product is chosen.
| Day | Job | Done when |
|---|---|---|
| 1 | Write the one-sentence purpose, name the owner, list the users | Written down and shared |
| 2 | Build stages and the ten fields | Owner can describe each stage in one line |
| 3 | Clean the data and import a test batch of ten | Ten records match the source |
| 4 | Import the live records and connect the website form | A test enquiry arrives with the right source |
| 5 | Run a thirty minute walk-through with everyone, on their own phones | Each person has logged one real action |
The following Monday, everyone works from the CRM and the old sheet is read-only. Do not run both.
Training that people use
Keep it short and tied to real jobs. A one-page sheet is better than a recorded course: “When you take a call, do this. After a site visit, do this. When you send a quote, do this.” Cover the mobile app on the same session, because that is where most updates will happen. The reasons staff drift away from a CRM, and the fixes, are in our guide to CRM adoption.
What to check at 30 and 90 days
Set two review dates now, before you are busy.
- At 30 days: Are new enquiries landing without typing? Does every open record have a next action and date? Which field does nobody fill in? Delete it.
- At 90 days: Which stage do records get stuck in? What are the top three lost reasons? Is the response time to a new enquiry shorter than before? Adjust stages and automations based on what you see.
Common ways implementation goes wrong
- Too many fields. Staff skip them, and then the data is patchy.
- No owner. Nobody decides, so nothing changes.
- Running two systems. The sheet and the CRM disagree and people trust neither.
- Automating before the process is clear. You send a follow-up to someone who already said yes.
- Training once. A new starter joins and learns a worse version from a colleague.
If you would rather have this done for you, CentriWeb does CRM set-up and looks after it afterwards. If your team is small and the tool you already have does the job, we will say so.