When Is It Time for Your Small Business to Get a CRM?
A plain decision guide for small businesses: when a spreadsheet is enough, the signs it is failing, a scorecard to fill in, and a migration checklist.

In this article
- What a CRM is, and what it is not
- Why a spreadsheet is often enough
- Signs the spreadsheet is failing
- Signs you are too early
- A scorecard you can fill in
- What to decide before you choose anything
- Choosing: what to ask, without naming winners
- A spreadsheet-to-CRM migration checklist
- Common mistakes
- Where software fits
- What to do this week
- Sources
Many small businesses do not need a CRM yet, and some never will. A sole contractor with a dozen active clients and a tidy spreadsheet can run a perfectly good sales process with nothing else. The useful question is not whether CRM software is good in general. It is whether the records you keep today are failing you in specific, nameable ways.
Spreadsheets do tend to fail in a predictable order. A follow-up gets forgotten. A second person starts editing the same sheet. Someone asks what is in the pipeline and the honest answer takes an afternoon. This guide is for the moment those things start to happen. You will get a plain definition of a CRM, the signs a spreadsheet has stopped working and the signs you are too early, a scorecard to fill in yourself, what to settle before choosing a tool, a migration checklist, and the usual ways the project goes wrong.
Good Loux publishes this Journal and sells software in this area, so a short section near the end says what our product does and does not do. Nothing before it depends on buying from us. The one worked example is hypothetical and labelled as such.
What a CRM is, and what it is not
CRM stands for customer relationship management. Vendors define it a little differently, and since this article is not about any one product, we cite two definitions only to show where the common ground is. Salesforce's own explainer describes CRM as "a technology for managing all your company's relationships and interactions with customers and potential customers." It says a CRM system stores customer and prospect contact details in one digital record, keeps the earlier interactions with your company on that record so staff can see the history, and helps teams identify sales opportunities and see their pipeline. Microsoft's documentation for Dynamics 365 Sales is more specific about the objects involved: it says you use the product to keep track of accounts and contacts and to nurture sales from lead to order, and that you can also create marketing lists and campaigns and follow service cases.
Strip the vendor language away and the core is three things, which is our own summary rather than either vendor's wording:
- A shared record of who. Companies, and the people at them, each stored once.
- A history of what happened. Calls, emails, meetings, quotes and notes, attached to the right company or person, in date order.
- A list of what happens next. For each open deal or relationship, the next step and the date it is due, visible to everyone who needs to see it.
That is the whole idea. A spreadsheet can hold all three, and that is why spreadsheets work for so long. What separates a CRM from a spreadsheet is not that it stores more. It is that it relates things to each other (this person belongs to that company, this note belongs to that deal), lets several people work in it at once without overwriting each other, and keeps a history instead of only the current state.
It helps to be equally clear about what a CRM is not.
- It is not a sales process. Software records a process, and does not invent one. If nobody has decided what happens after a first call, the CRM will faithfully record that nothing happens.
- It is not a source of leads. A CRM holds the people you already know about. Finding new ones is a separate job; we wrote about it in How to Build a Prospect List From Scratch.
- It is not, at its core, an accounting or invoicing system. The same Microsoft page shows how far one large product can stretch, into marketing campaigns, service cases, forecasting and AI features. In our view a small business can pay for, and be confused by, a great deal it will not use.
- It is not a substitute for talking to customers. A record that says "interested" tells you what someone wrote down, and not what the customer thinks.
Why a spreadsheet is often enough
A spreadsheet has three real advantages for a small business, and they deserve to be named before we talk about its failures.
It costs nothing you are not already paying for. It takes ten minutes to change a column. Everyone already knows how to use it. And while you are still working out what you need to record, being able to add a column on Tuesday and delete it on Friday is worth a lot, because most small businesses do not know their real process until they have run it for a few months.
A spreadsheet also forces a certain honesty. If your sheet has forty rows and you can see all of them on one screen, you know your pipeline. The trouble starts when the sheet stops being something one person can hold in their head.
For one person, with roughly a dozen live opportunities, a sheet with these columns is hard to beat:
Company Person and role How to reach them Where they came from Stage (new, talked, quote sent, waiting, won, lost) Last contact date Next step Next step due date Notes (one line, dated)
If you keep it this plain and you open it every working day, you may not need anything else yet. The most important column is the next-step date. A sheet sorted by that column, with overdue dates highlighted, is a primitive but effective follow-up system.
Signs the spreadsheet is failing
These are the symptoms we think matter, in rough order of how soon they appear. This list is our own analysis, drawn from common small-business experience, not from a study, and none of the signs is a rule. Two or three of them together is a stronger signal than any one alone.
Follow-ups are being missed. A quote went out three weeks ago and nobody has looked at it since. The sheet has a "next step" column but nothing prompts anyone to read it. This is the earliest and most expensive sign, because a missed follow-up is a lost sale that nobody notices as a loss.
Two or more people edit the sheet. A shared file solves the "which copy is current" problem only partly. Two people editing the same row at the same time overwrite each other. People keep private copies "just for my own accounts". Formatting drifts. Somebody sorts one column without the others and scrambles a hundred rows, and the damage is found a week later.
There is no history. The sheet shows where a deal is now and not how it got there. A note typed over the old note erases the old note. When a customer says "I told your colleague about this in March," nobody can check.
You cannot answer "what is in the pipeline?" quickly. If a partner, a lender or your own planning needs the number of open deals by stage and the total value, and producing it means filtering, copying and adding by hand, the sheet is no longer a working tool. It is an archive that you have to process.
Contact information lives in several places. Some of it is in the sheet, some in an email inbox, some in a phone's contacts, some on a business card in a drawer. The same person appears twice, spelled two ways, with two different phone numbers.
Someone leaves and the records go with them. A salesperson, an office manager or a subcontractor who handled relationships leaves, and what they knew was in their head, their personal inbox or a private sheet. This sign is the one to take most seriously, because it is the only one that can cost you customers outright and in a single week.
A less obvious sign: you have started avoiding the sheet. If people stop updating it because it is tedious, then the record is already out of date, and the problem is not the discipline of your staff. Either the tool or the process is wrong.
Signs you are too early
Being too early is a real and common mistake. A CRM bought before you have a repeatable way of selling becomes an expensive way to record confusion.
You have fewer than a handful of active deals. If you have three open quotes, a spreadsheet, a calendar and a good memory are enough. There is no fixed threshold, and we are not going to invent one. The useful test is whether you can list every live opportunity and its next step from memory in a couple of minutes. If you can, you do not have a record-keeping problem yet.
You have no repeatable process. Every sale has gone differently, and you could not describe the stages a typical deal moves through. A CRM asks you to define stages. Define them wrongly and you will spend months fighting the tool. It is cheaper to run a spreadsheet for a quarter, note what actually happens between first contact and signed agreement, and then encode that.
You are unclear who you sell to. If you are still trying three different kinds of customer to see which one works, the records you want are about experiments and not relationships. A sheet with a "segment" column serves that better than a system with a rigid structure. The work to do first is described in How to Find Your First 10 B2B Customers.
You expect the software to create discipline. If the real issue is that nobody makes follow-up calls, no tool will change that. A weekly twenty-minute review of the pipeline, run by one named person, will do more than any purchase.
You are one person who already has a system that works. A sole operator whose records are complete and whose sales are steady does not have a problem to solve. Change when something hurts.
A scorecard you can fill in
The scorecard below is our own. It is a way of making your judgment visible, not a validated instrument, and nothing in it comes from a study. Score each line 0, 1 or 2 for your business as it is now, not as you hope it will be. Do it alone first, and if more than one person sells, have each of them do it separately and compare.
Part A: Pain (how much is the spreadsheet hurting?)
A1. Follow-ups missed or late in the last 90 days:
0 = none that I know of, 1 = once or twice, 2 = several, or I cannot tell
A2. People editing the same records:
0 = one person only, 1 = two people, occasionally, 2 = several people, or private copies exist
A3. History of a customer (who said what, when):
0 = I can reconstruct it easily, 1 = with effort, 2 = I often cannot
A4. "What is in the pipeline?" (count and value by stage):
0 = under five minutes, 1 = under an hour, 2 = longer, or I avoid asking
A5. Contact details in several places:
0 = one place, 1 = two places, 2 = three or more
A6. Dependence on one person's memory or private inbox:
0 = low, 1 = moderate, 2 = if they left, we would lose accounts
Part B: Readiness (are you ready to use one properly?)
B1. Can you name the stages a typical deal passes through?
0 = no, 1 = roughly, 2 = yes, and you have used them for a few months
B2. Do you know who you sell to?
0 = still experimenting, 1 = mostly, 2 = yes, with a written description
B3. Is there a named person who will own the system and keep it clean?
0 = no, 1 = someone could, 2 = yes, with time set aside
B4. How many deals are open at once?
0 = fewer than five, 1 = five to fifteen, 2 = more than fifteen
B5. Can you say what you must record, in ten fields or fewer?
0 = no, 1 = roughly, 2 = yes
Add Part A (0 to 12) and Part B (0 to 10).How to read it, in our judgment:
- Part A low (0 to 3), Part B anything: stay on the spreadsheet. Spend the effort on the weekly review.
- Part A high (7 or more), Part B high (6 or more): you are probably ready, and you are probably late.
- Part A high, Part B low: the pain is real, but the process is not ready. Fix the process first, in the spreadsheet if need be: define the stages, name the owner, trim the fields. You will be able to choose a tool far better after that.
- Part A middle (4 to 6): try cheaper fixes first. A due-date column with conditional formatting, a single shared file with one editing rule, and a fifteen-minute Monday review fix a surprising share of the problems on the list.
The cutoffs are a rule of thumb, not a calibrated threshold. If your totals sit on a boundary, trust the individual lines: a 2 on A6 alone (a departure would cost you accounts) can justify a move that a total score would not.
What to decide before you choose anything
In our experience, a CRM project is won or lost before you pick the software. Settle these before you open a vendor's website.
Who will use it. List the people, and be specific about what each does in it: enters records, reads them, reports from them. A tool for one owner-operator and a tool for ten salespeople are different purchases. Ask each person what they need to see on a Monday morning. A system designed around what the owner wants to read, and not what the people entering data find easy, will be filled in badly.
What must be recorded. Write the list of fields, and keep it short. For most small businesses that serve other businesses it looks like our minimal sheet above plus a few more: company, person and role, source, stage, value, next step and date, and notes. Everything else is optional until you can name the decision it supports. The common mistakes below come back to this.
How the data gets in. Know where your records are today (the sheet, email, a phone) and who will clean them. Microsoft's import documentation for its customer engagement apps gives a useful picture of what any import involves: you prepare source files, in formats such as CSV, and you must map every column in the file to a field in the system, because, per that page, unmapped data is not imported. It also describes a duplicate-detection option that can be enabled to prevent importing duplicate records. That is one vendor's tool, and others differ, but the structure is general: your columns have to line up with the new system's fields, and duplicates are a known hazard. Both are far easier to handle if your spreadsheet is consistent before you start.
What switching away would cost. People tend to evaluate the move in and forget the move out. In two years you may want a different tool, or a cheaper plan, or you may stop. Ask in advance what that costs in time, in data and in lost history. The answer depends partly on the next point.
Who owns the data, and whether you can export it. The customer records belong to your business, and you should be able to leave with them. Before you commit, find the vendor's own documentation on exporting data, find out in which formats it exports, and whether notes and history are included or only contact fields. We have not surveyed vendors' export features for this article, and we are not going to assert how any particular one works. Check it yourself, in writing, for each product you consider, and test an export during any free trial.
Privacy and data protection obligations. A CRM holds personal information about named people: names, work emails, phone numbers, notes about them. What you owe the people on your list depends on where you and they are, what you collect and what you do with it, and this is general information, not legal advice. Our one cited source is the Federal Trade Commission's guide, Protecting Personal Information: A Guide for Business (October 2016), which is written for U.S. businesses and treats the matter as basic data hygiene. It lists five principles. "Take stock": know what personal information you have in your files and on your computers. "Scale down": in its words, "If you don't have a legitimate business need for sensitive personally identifying information, don't keep it." It recommends least privilege, so each employee reaches only what the job requires. "Lock it": require strong passwords, consider multi-factor authentication, and make sure people who leave or move to another part of the company no longer have access to sensitive information. "Pitch it": the guide's point is that discarded records are still a risk, and it says that what looks like a sack of trash to you can be a gold mine for an identity thief. "Plan ahead": have a plan in place to respond to security incidents, and consult your attorney about notification. For a sales CRM, the practical takeaways are modest: do not store what you do not need, give people access only to what they use, remove access the day someone leaves, and know where the data sits. Rules that apply to the messages you send from that data, including opt-outs, are covered in the prospect-list article linked above. We did not research state privacy statutes or rules outside the United States, and they can differ materially, so check yours.
Choosing: what to ask, without naming winners
This article does not recommend a product, and we have not tested any of them. What we can offer is a list of questions that apply to any tool in the main categories: a general-purpose contact-and-deal CRM, an industry-specific system for a single trade, a pipeline tool built around sales stages, or a spreadsheet-like database that has been shaped into a CRM. You will find that the categories overlap, and a product's marketing page is a poor guide to what it actually does.
Ask:
- Can I see every company, its people, its conversations and its next step on one screen?
- Can two people work at once without overwriting each other, and is there a history of changes?
- Can I get a report of open deals by stage and by value without building it myself?
- What does the import look like, and what does it do with duplicates?
- Can I export everything, including notes and history, in a plain format, and have I tried it?
- Can I set access by person, and remove it in one action?
- What does it do that I do not need? Every unused feature is something to learn, configure or ignore.
- If I stopped paying, what happens to my data, and how long do I have to take it out?
Fill a free trial with twenty of your real records and run a week of actual work through it before you pay for anything. The trial is a test of whether your people will use it, which is the question that decides whether it pays off.
A spreadsheet-to-CRM migration checklist
If the scorecard says go, the migration is a small project with a clear order. Do the cleaning in the spreadsheet, where it is easy, and not in the new system, where it is not.
Before [ ] Name one owner for the system and set aside regular time for them (an hour a week is realistic at first). [ ] Write down the stages of a typical deal, in the words your people use. [ ] Write the list of fields. Ten or fewer to begin with. For each, note who fills it in and what decision it supports. [ ] Back up the current spreadsheet and any other source files, and keep the copies read-only. [ ] Decide which records to bring over (open deals, active customers, people you plan to contact) and which to archive. Clean (in the spreadsheet) [ ] One company per row, one spelling per company. [ ] Merge duplicate companies and people. Choose the surviving record and keep any notes from the others. [ ] Put dates in one format. Put phone numbers in one format. [ ] Split fields that hold two things (for example "Jane Smith, owner" in a single cell). [ ] Remove records you could not defend contacting. Mark anyone who asked not to be contacted, and keep that mark through the move. [ ] Gather contacts that live in inboxes, phones and business cards into the sheet. Import [ ] Map each column to a field. Check what the tool does with columns that do not map. [ ] Import a small test batch (twenty rows), open the records and compare them with the sheet. [ ] Fix the mapping, then import the rest. [ ] Spot-check fifty records by hand, including a few of the messiest ones. Run in parallel [ ] For two to four weeks, use the new system for all new activity. Keep the old sheet read-only for reference. [ ] Hold the weekly pipeline review from the new system, not from the sheet. [ ] Ask each user what they dislike. Change a field or stage if the same complaint comes from two people. After [ ] Test a full export and open the file. [ ] Set up access by person, and write down who removes access when someone leaves. [ ] Retire the old sheet only when nobody has opened it for a month.
Two notes on that list. The "run in parallel" stage is where migrations succeed or fail, because the risk is not the import but the habit: people go on updating the old sheet because it is familiar. A single rule, stated plainly (the sheet is read-only from day one of the trial), prevents most of that. And the "defend contacting" step is deliberate. A migration is a rare moment when you look at every record, so use it to drop the ones that should not be there.
Common mistakes
Buying for the features. A feature list is a list of things the vendor can show in a demo. You need the three things from the start of this article: who, what happened, what is next. Evaluate those first, and treat everything else as optional until you can name the problem it solves for you. The Microsoft page we cited lists campaigns, service cases, forecasting and AI features in a single product, which shows the breadth that is available, and does not say you need any of it.
No owner. A system nobody owns drifts. Fields are used inconsistently, duplicates accumulate, stages mean different things to different people, and within a quarter it is as unreliable as the spreadsheet it replaced. One named person should own the structure and the tidiness, even if everyone enters the data.
Too many fields. Every field you add is a small tax on every record, paid by whoever enters it, forever. Fifteen blank fields teach people to leave things blank. Start with ten or fewer, and add a field only when someone can name the decision it will change.
Importing messy data. A CRM will not clean your data. It will store the mess more permanently and make it look more official. Duplicates become harder to find when they are spread over thousands of records, and a wrong phone number looks just as authoritative in a database as in a sheet. Spend the clean-up time before the import.
Skipping the review. The software shows what is overdue, but it does not make anyone act. Hold a short weekly pipeline review, with a fixed agenda: what moved, what is stuck, what is overdue, what is next. If you want a structure, our guide to building a sales pipeline from scratch covers stages and the review in more depth. Retention matters too: a record of what customers have bought and when they last heard from you is what makes the work in How to Retain Commercial Cleaning Clients possible, and the same logic applies in any trade.
Treating the choice as permanent. You will probably change tools at least once. If your records are clean and exportable, that is a weekend and not a crisis. The article on the systems every small business needs before it starts growing puts the CRM question in the context of the other systems a growing company needs, which are usually more pressing than software for sales.
Where software fits
Good Loux is ours, so weigh this section accordingly. We sell a business development and CRM platform, and we are not neutral about it.
Whichever way you go, the test is the same one the scorecard applies. The tool is worth having when it fixes a named failure in your records, and not before.
What to do this week
- Fill in the scorecard, alone, and then with anyone else who sells. Note where your answers differ.
- Add the next-step date column to your sheet if it lacks one, sort by it, and highlight overdue rows.
- Hold one pipeline review, twenty minutes, with the agenda: moved, stuck, overdue, next.
- If Part A is high and Part B is low, define stages, name an owner and list your fields, in the spreadsheet.
- If both are high, write your list of ten fields, answer the eight questions above for two or three tools, and trial each with twenty real records.
- Test an export from anything you try, before you put real data in.
Sources
We read these directly on November 19, 2026. Pages change, so check the originals.
- Salesforce: What Is CRM (Customer Relationship Management)? for how one vendor defines CRM and what it says a CRM system stores (contact records, interaction history, pipeline). Used as a definition only, not for any statistic or claim about results.
- Microsoft Learn: Welcome to Dynamics 365 Sales (page dated June 2026) for how one vendor's documentation describes accounts, contacts, the lead-to-order process, and the range of features, including marketing campaigns and service cases, in one product. Used as a definition and a description of scope only.
- Microsoft Learn: Import data (all record types) for how one vendor's import works: source files in formats such as CSV, mapping every column to a field with unmapped data not imported, and an optional duplicate-detection setting. Used as an example of what any import involves, not as a statement about other products.
- Federal Trade Commission: Protecting Personal Information: A Guide for Business (October 2016) for the five principles and the quoted sentence on keeping only what you need, plus the passwords, access and response-plan points. General information, not legal advice.
We cite no adoption, cost or return-on-investment figures, because we did not find a primary government or standards-body source for them, and vendor statistics are marketing. The scorecard, its cutoffs, the signs of failure and of being too early, the checklists, the questions to ask and the hypothetical HVAC example with its arithmetic are our own analysis and opinion, not research findings. Everything about how Good Loux works comes from the product itself.


