How to Choose a CRM in Israel: The Checklist That Decides Before Price Does

Choosing a CRM in Israel almost always turns on the same four things: Hebrew and RTL that genuinely work, a connection to whatever issues your invoices, WhatsApp, and the ability to leave with your data. Check those four before looking at price and the decision holds for years.

The four critical checks

The last row is the only one that is easy to verify before signing and very hard after.

The twelve questions

1. How a name, address and document look in Hebrew - on screen, in exports and in outgoing email.
2. Which integration exists for your invoicing system, by name.
3. Where a WhatsApp conversation is stored, and who can see it.
4. How you search for a customer by partial phone number or a misspelled name.
5. What happens when an existing lead comes back - new record or existing one.
6. Which fields you can add yourself, and which need the vendor.
7. What permissions look like: who can *see*, not only who can edit.
8. What the system sends automatically, and how you stop a send.
9. What happens when a rep leaves - who inherits the customers and the history.
10. What is included and what costs extra: users, storage, automations, support.
11. What a full export looks like, and in what format.
12. Who configures the system on your side, and for how many hours a month.

Question 12 predicts success better than any other, and usually has no answer in the first meeting.

Why the pricing model changes the decision

Per-user pricing sounds simple until you see what it does to behaviour: businesses start sharing a login to save money, and from that moment you cannot tell who did what. That destroys the one thing such a system is meant to deliver - clear ownership and a record.

So it is worth asking three questions around pricing: whether a low-cost view-only user exists, what happens to price when record counts double, and what happens when you want to add a user for a one-month trial. A vendor with a reasonable answer to all three lets you grow without inventing workarounds.

What usually breaks a year later

• The invoicing connection that was "in development" and stayed there.
• Automated messages still going out to customers who are no longer relevant.
• Duplicate fields created because nobody tidied up, with no way to know which to use.
• A report showing a number nobody can explain.
• A departed rep whose customers are still assigned to them.

Each one is prevented by a single question from the list above, and each costs working days when discovered late.

The evaluation order that saves time

1. Write the process in seven steps before looking at any system.
2. Filter vendors on two checks only: invoicing and RTL.
3. Run a demo on your process, not a generic demo.
4. Ask for a sample export and open it.
5. Pilot with one user and five real customers.
6. Only then discuss price.

This order looks slow and is faster: it eliminates two or three vendors at step 2, and prevents the common outcome of deciding on price and then discovering a critical connection is missing.

What is not worth evaluating

• Long feature lists. Nearly identical across every system.
• Screen design. You get used to it within a week.
• AI as a general promise, with no concrete example on your own data.
• The vendor's customer count. It says nothing about fit to your process.

What is worth the time is speaking to two of the vendor's existing customers and asking one question: what surprised you after three months.

Why does the RTL check matter more than it seems?

Hebrew in a CRM breaks in three separate places, and usually only the first is checked. On screen - nearly everyone is fine here. In exports - a file opened in Excel with the wrong encoding turns names into gibberish, and it surfaces only when you have to hand a list to an accountant or another system. In outgoing messages - an email or SMS where a phone number or document number flips inside a Hebrew sentence, and the customer receives wrong text.

The test takes ten minutes: enter a customer with a Hebrew name including gershayim in the business name, an address with a house number, and a long description; send yourself an outgoing message; then export the record to a file and open it. If all three look right, the system will probably handle Hebrew in the places you did not test as well.

What must be retained per customer

Before comparing systems, decide what must be retained about each customer for the next three years: contact details, lead source, what was offered and at what price, what was agreed, what was sent and when, and who spoke to them last. That list is a specification - and any system with no natural place for it will generate a side spreadsheet to fill the gap.

It is also the most efficient way to shorten demos: instead of asking "what do you have", ask each vendor to show exactly where those six items are stored. The difference between systems becomes clear in five minutes, and it is often nothing like the impression their website gives. If the answer for several of them is "you can add a custom field", that is fine - but ask who adds it and whether it is searchable, because a note field nobody can search is not a record. Where no system fits the list without workarounds, the honest comparison becomes custom CRM versus off-the-shelf.

Sources

• Fireberry by Powerlink
• Kala CRM
• monday.com