Most companies that switch WhatsApp CRM within the first year didn’t choose a bad tool. They chose a tool their team didn’t use. These are different problems, and the second one doesn’t appear in any feature comparison chart — but it’s what determines whether the investment pays off or just becomes a dormant monthly fee.
In this article
Before the criteria: write down your process
There is one mistake that precedes all others: comparing tools before knowing what the company actually does. Without this, you end up choosing based on the prettiest demo or the most persistent salesperson.
Take a sheet of paper and answer in five minutes:
- Where do your contacts come from today? Ads, referrals, Google, physical store, Instagram. List the channels and the approximate weight of each.
- What are the steps until the sale? Write the real ones, not the ideal. If today it’s "contact → send price → client disappears → sometimes returns," write that down.
- Who does what? Does one person do everything or is there screening and then a specialist?
- What is lost the most? Contact without response, proposal without follow-up, old client never contacted again.
This sheet becomes the evaluation script. Every demo then has an objective question: how does this tool solve exactly this?
The 9 criteria, with weights
1. Type of connection — high weight
Check if the tool offers QR Code and Meta’s official API. Not because you will use both today, but because today’s choice shouldn’t block tomorrow’s. Those who only offer QR Code leave you stuck when the volume of active messages grows. Those who only offer API force you to migrate the number right away and pay per message from day one, even with a small operation.
2. Billing model and team quota — high weight
Two numbers decide: how many people are included and how much each additional person costs. A lower monthly fee with high per-user charges gets expensive fast — and worse, it incentivizes not giving access to those who need it. Always simulate with the team size you expect in twelve months.
3. Editable funnel — high weight
The stages need to be yours. If the tool imposes a fixed funnel, one of two things happens: either the team forces their process into a wrong mold, or simply ignores the funnel and it becomes decoration. Ask during the demo: can I create, rename, reorder, and delete stages myself without calling support?
4. Conversation distribution and assignment — high weight
Each conversation needs an owner, and the tool must make that clear. Check if there is rotation or queue, if you can transfer with history intact, if the system notifies when another agent is already responding and — an item almost everyone forgets — if a returning client automatically goes back to the agent who previously attended them.
5. Automation and broadcasting — medium weight
Here the feature list can be misleading. What matters: automatic message outside business hours, follow-up reminders for stalled conversations, automatic distribution by origin, and segmented sending to lists. For broadcasting, one specific question separates serious tools from risky ones: does it check if the number has WhatsApp before sending? Sending to nonexistent numbers is the fastest way to accumulate delivery errors and draw negative attention to your number.
6. Permissions and database ownership — high weight
A salesperson should see their own portfolio, not the entire database. And exporting contacts should be an administrator privilege. If any agent can download the entire database in a file, your client portfolio is one resignation away from becoming a competitor’s prospect list.
7. Reports that answer management questions — medium weight
A pretty chart is not a report. The test is whether the tool answers, without exporting anything: how many new conversations entered in the period, average time to first response, how many opportunities are open and their total value, conversion by agent and by origin. If you need to download spreadsheets and cross data manually to answer this, no one will do it after the second month.
8. Export and exit — medium weight
Ask directly: if I want to leave in two years, what do I take and in what format? The ideal answer includes contacts, custom fields, conversation history, and funnel opportunities, in an open format. A vendor who hesitates on this question is saying something important.
9. Time to get running and support — medium weight
A tool that requires weeks of implementation is usually too big for a company of five to twenty people. Also confirm support in writing: channel, hours, and response time. When the company’s number stops responding on a Monday morning, email support with a three-business-day response time is not support.
The tenth criterion, not in the table: how many clicks the team needs to perform the day’s most repeated task. If replying to a client and moving them to the next stage costs six clicks instead of two, the tool will be abandoned in eight weeks — regardless of all the good things it does.
Scoring table to compare proposals
Give a score from 0 to 3 for each criterion for each vendor, multiply by the weight, and sum. It’s deliberately rudimentary: it forces you to compare the same things and removes the decision from the realm of impression.
| Criterion | Weight | Vendor A | Vendor B | Vendor C |
|---|---|---|---|---|
| Connection (QR + API) | 3 | |||
| Billing and team quota | 3 | |||
| Editable funnel | 3 | |||
| Distribution and assignment | 3 | |||
| Permissions and database | 3 | |||
| Automation and broadcasting | 2 | |||
| Reports | 2 | |||
| Export and exit | 2 | |||
| Implementation time and support | 2 | |||
| Total (maximum 69) | — |
One note on how to read the result: if the difference between two vendors is less than 10%, consider it a technical tie and break the tie with the test in the next section — not by price. Price should only decide between truly tied options.
Continue reading for free
Fill out once and unlock this and all other blog articles. We don’t charge anything — we just want to know who we’re writing for.
Takes 20 seconds. Prefer to talk directly on WhatsApp?
The 30-minute test
A guided demo shows the tool on its best day. This script shows how it behaves on yours. Request access and do it yourself, with a real number:
- Connect the number (5 min). Time it. If it takes more than fifteen minutes and requires support, that says a lot about the rest.
- Build your funnel (5 min). Use the stages from the sheet you wrote at the start. If you can’t do it alone, that’s a problem — because you’ll want to change this again in three months.
- Receive a real conversation (5 min). Ask someone to write to the number. See how long it takes to appear on the dashboard and if the notification works.
- Go through the entire funnel (5 min). Move the opportunity from stage to stage until closing. Count the clicks.
- Transfer to another agent (5 min). With a second login. Confirm that the history transferred and the other agent sees everything they need.
- Write an internal note and a report (5 min). The note must be invisible to the client. And try to answer, just by looking at the screen: how many opportunities are open and what is their total value?
If any of these six steps cannot be done in the test — because it depends on support configuration, a higher plan, or an implementation meeting — note it. This is information about what your daily life will be like.
Warning signs during the demo
- They don’t let you interact. A demo where only the salesperson controls the screen hides something. Ask for the keyboard.
- They promise "AI solves it." Automation with artificial intelligence really helps with screening and answering repeated questions. But if the answer to every process problem is "AI takes care of it," the process still has no owner.
- Price only after the meeting. Not always bad faith — complex operations require scope. But for a small or medium company, a public price list is a sign there won’t be surprises.
- Result numbers without origin. "Our clients sell 40% more." Ask how many clients, in what period, measured how. Without a specific answer, it’s just decoration.
- Contract only at signing. Ask for the contract and cancellation policy before deciding. Loyalty and penalties need to be accounted for from the start.
The 12 questions to send in writing
A demo is a conversation, and conversations fade. Before deciding, send these twelve questions by WhatsApp or email to each vendor in the running and keep the answers. Two things happen: you compare apples to apples, and you have a record of what was promised.
- Does the tool connect via QR Code, Meta’s official API, or both?
- How many people does the plan include and how much does each additional person cost per month?
- How many WhatsApp numbers does the plan include and how much does each additional number cost?
- Is there an implementation, training, or setup fee? How much?
- Is there a loyalty period? What is the penalty for early cancellation?
- What is the full monthly fee after any introductory promotion?
- Can I create and reorder funnel stages myself without opening a support ticket?
- How does conversation distribution among agents work? Does a returning client go back to the agent who attended them?
- Does mass broadcasting check if the number has WhatsApp before sending?
- Can a regular agent export the contact database? Can this be restricted?
- If I cancel, what can I export — contacts, conversation history, funnel — and in what format?
- What is the support channel, hours, and response time for the plan I’m evaluating?
Note that none of these ask about pretty features. They are all about real cost, autonomy, and exit — the three areas where surprises usually appear after signing. If one vendor answers nine of the twelve clearly and another answers five evasively, you already have a better tiebreaker than any price difference.
Three mistakes that lead to switching in less than a year
Choosing by feature list. Every mature tool has a long list, and almost no one uses more than 20% of it. Comparing lists rewards whoever writes more items, not whoever solves better. Compare by what you identified in your process sheet, and ignore the rest.
Deciding alone, in the office. The team using the tool eight hours a day is the sales team. If the manager chooses and imposes it, resistance appears in the third week disguised as "the system is slow." Include at least one salesperson in the 30-minute test and take their opinion seriously — they will notice frictions you don’t.
Migrating everything at once. Trying to import five years of contacts, activate all automations, and train ten people in the same week is a recipe for the team to go back to the old phone. Start with open deals, one automation only, and one agent for one day. The rest will come in by itself.
Frequently asked questions
What is the most important criterion?
Team adoption. A tool the team doesn’t open every day has zero return, no matter how good the feature list is. In practice, this translates into three things: interface response speed, funnel that mirrors the company’s real process, and few clicks for repeated tasks.
Do I need Meta’s official API from the start?
Not necessarily, but choose a tool that supports both paths. Starting with QR Code is faster and has no per-message cost; if the volume of messages initiated by the company grows, you migrate without changing vendors.
Does a generic CRM tool work for WhatsApp?
It works for storing contacts and deals, but usually treats WhatsApp as an attachment: the conversation happens outside and someone needs to record it later — which no one does. If WhatsApp is your main sales channel, the conversation needs to be inside the system, not beside it.
How long until I see results?
Organization appears in the first week: you start to see how many conversations are stalled. Sales results depend on the team using the funnel disciplinarily, which usually takes thirty to sixty days — and depends much more on internal agreement than on the tool.
Try the 30-minute test on VeyloCRM
Connection via QR Code or Meta’s official API, editable funnel, conversation distribution, and broadcasting with number check — all on the same dashboard.