Private School CRM: Do You Need One, or Will a Sheet and AI Do?

Most small private schools do not need a CRM until they handle roughly 150 enquiries a year or share admissions between two or more people. Below that, a well-built sheet with AI-drafted follow-ups does the job for free. Above it, missed follow-ups start to cost places, and a CRM earns its keep.

The catch is that the number is a symptom, not the cause. What breaks a sheet is not volume alone but the number of people editing it, the number of steps between enquiry and a place, and how many weeks a family takes to decide. A school with 90 enquiries and three staff touching them can need a CRM sooner than one with 160 and a single admissions officer.

Follow me on Instagram@sagnikteaches

What a school actually needs from a CRM

Strip away the product language and an admissions system has to do five things: remember every family who asked, show who needs a reply today, keep the history of what was said, move each family through the stages from enquiry to enrolment, and let you report on where places come from. Anything beyond that is nice to have.

Connect on LinkedInSagnik Bhattacharya

A sheet can do all five at small scale. It cannot do them when several people are in it at once, when you need a reminder that fires without anyone opening the file, or when you want to send a sequence of emails and see who opened them. Those three gaps are the real reasons to move. If you cannot name which gap hurts, you probably do not need a CRM yet. The tutorial on outgrowing spreadsheets lists more warning signs.

Subscribe on YouTube@sagnikteaches

A worked comparison for a school with 180 enquiries a year

Take an illustrative small independent school: 180 enquiries a year, one admissions officer at two days a week, a head who drops in for tours, and an open morning each term. Each enquiry needs roughly five touches before a decision: a first reply, a prospectus, a tour offer, a reminder before the tour and a post-tour follow-up.

That is 900 touches a year, about 17 a week. At five minutes each to find the family, write the message and log it, that is roughly 85 minutes a week. The question is how much of that a tool removes, and what the cost of a missed follow-up is.

HubSpot's free CRM covers up to two users and 1,000 contacts, with no time limit. Its Starter tier lists at $20 a seat a month; the pricing page currently shows a much lower introductory price for new customers, so plan with the list price. Many of HubSpot's AI features and agents run on HubSpot Credits: Starter includes 500 a month, and extra capacity costs $10 per pack of 1,000. School-specific products rarely publish a price, so I will not guess one; ask for a written quote with the contract length, set-up fee and what happens to your data if you leave.

On the numbers above, the sheet costs nothing in licence and perhaps 85 minutes a week of the admissions officer's time. HubSpot free costs the same licence and saves some logging, but you spend a day or two setting it up. A school-specific tool might save more time but at a price you cannot assess until you ask. The decision rule: stay on the sheet until two of the three gaps (several editors, no automatic reminders, no email sequences) bite, then try the free CRM before you ask for a school-specific quote.

Building a sheet that behaves like a small CRM

A good enquiry sheet has one row per family, not per child, with children in a separate tab or a numbered column set. The columns that earn their place:

• Family reference (a number, not a name) and a contact name
• Email, phone, preferred contact method
• Child's age or year group at entry, and intended start term
• Source (website, open morning, referral, event)
• Stage: Enquiry, Prospectus sent, Tour booked, Toured, Application, Offer, Enrolled, Declined, Lost
• Last contact date and next action date
• Consent to contact, and the date given
• Delete-by date

Two formulas turn it into a to-do list. Assuming next action date is in column L, this flags what is due today or overdue:

=IF(AND(L2<>"", L2<=TODAY()), "DO TODAY", "")

And this counts days since last contact, which you can sort by to find families who have gone quiet:

=IF(K2="", "", TODAY()-K2)

Add a filter view called "Due now" that shows only rows marked DO TODAY and you have the daily list a CRM would give you. For writing the formulas from plain English, the Google Sheets AI prompts tutorial has ready-made requests. The sheet is also where an AI function can help, if your Google plan includes it. Google Sheets has an AI function, written =AI("prompt", range), that runs a prompt over a selection of cells, and only the first 350 selected cells generate at a time. It needs an eligible Google Workspace or Google AI plan, so check yours first, and use it only on columns that hold no personal detail, such as tidying the free-text source column into your fixed categories.

AI follow-ups that sound like a person wrote them

The biggest time saving for a single admissions officer is not the sheet but the first draft of each message. The safe pattern is templates with placeholders, filled by you. A prompt that works:

Write a follow-up email from a small independent school's admissions office. Situation: a family toured on Tuesday, their child is joining in September, we have two places left in that year group. Tone: warm, not pushy, plain English, under 120 words. Include: thanks for visiting, one specific thing to mention [detail from tour], the next step (application form, deadline [date]), and an offer to answer questions by phone. Use [first name] and [child's first name] as placeholders. Do not promise a place.

An illustrative output: "Dear [first name], thank you for coming in on Tuesday. It was lovely to meet [child's first name] and I hope the classroom visit gave a good feel for the school. You mentioned [detail from tour], so I have attached a note about it. Our application form is on the website and the deadline for September is [date]. If it would help to talk anything through, I am happy to ring at a time to suit you." What to fix: the AI tends to write "limited places remaining" as pressure. If you want to avoid that, say so. Remove any claim you cannot back, and add the real detail from the tour, which is the one thing the AI cannot know.

Three more message types worth drafting

The quiet-family nudge. Sent when there has been no contact for 21 days. Short, no pressure, one question: is there anything you would like to know? Keep it to three sentences.

The no-longer-interested close. A courteous message after the final nudge, saying you will keep their details for a stated period unless they ask you to delete them. This keeps the data tidy and the relationship kind.

The open-morning reminder. Sent 48 hours before, with the start time, where to park and who will greet them. The same reminder logic appears in the parents' evening tutorial, and the wording carries over.

Data protection with children's information

This is the part that deserves more care than the tool choice. An admissions record is personal data about adults and about children, and some of it can be sensitive: medical needs, learning support, family circumstances, previous school reports. Data-protection law treats children's data with extra care, and a school has a duty to its families that goes beyond the legal minimum.

1. Collect less. At enquiry, you need contact details, the child's age and intended start. Medical and learning-support information belongs later, in the application, in the school's own records.
2. Keep it out of AI prompts. A drafting prompt needs a template and placeholders, not a child's file. If you ever do need an AI to read real records, do it inside a tool with a data-processing agreement and a no-training setting, and write that decision down.
3. Control access. In a sheet, share with named people only, never "anyone with the link". In a CRM, give each person the lowest role that works.
4. Set retention. A delete-by date on every row, and a calendar reminder each term to clear them. Families who enrol move into the school's main records, so the enquiry copy can go.
5. Say it on the form. A short notice at the point of enquiry: what you collect, why, how long you keep it, how to ask for deletion.

If your school has a data-protection lead or governor responsible for it, show them the sheet or CRM design before launch, not after. For the tool side, the vendor questions tutorial lists what to ask any supplier that will hold family information.

Three schools, three different answers

A nursery-to-age-eleven school with 60 enquiries a year

One head, one part-time administrator, tours by appointment. At this size a sheet is not just adequate, it is better: a CRM would take longer to learn than the whole admissions year takes to run. Add the "Due now" view, a delete-by column and a weekly ten-minute review on Monday, and stop there. The risk is not capacity but forgetting, so put the review in the diary as a recurring event.

A senior school with 400 enquiries and a registrar plus two assistants

Three people in one file is the failure case. Even with careful colour codes someone will overwrite a row during an open evening, and nobody will know. Here a paid CRM tier is the minimum, because HubSpot's free CRM stops at two users, and a school-specific tool is worth pricing seriously, particularly if it links to the school's records so an accepted family does not have to be keyed in twice. The AI contribution stays the same: drafts and summaries, reviewed by a person.

A two-campus school with a shared admissions inbox

The volume may be moderate, but two campuses mean two tour schedules and two sets of places. Add a campus column to the sheet first and see whether filtered views solve it. If the admissions inbox is the real pain, a shared-inbox tool or CRM email logging may help more than a full system. If each campus answers its own enquiries, two separate sheets with an identical layout may be simpler than one combined one.

A quick sum to test your own case: take last year's enquiries, multiply by touches per family, multiply by five minutes, divide by 60. If the result is under 50 hours a year, no tool will repay its set-up. Over 150 hours, an hour saved per week is worth real money. Between, the data-protection and shared-editing arguments decide.

When a CRM starts to pay for itself

Four signals, in rough order of how reliable they are:

• A family got two different replies. Two people answered the same enquiry because neither knew the other had. One embarrassment like this is enough to take seriously.
• Follow-ups depend on someone's memory. If the admissions officer is off for a week and enquiries sit unread, the process is a person, not a system.
• You want to see source performance. "Which open morning produced applications?" is hard in a sheet and one chart in a CRM.
• Volume crosses the threshold. Around 150 to 200 enquiries a year with several touches each is where manual logging stops being sustainable.

An illustrative example of the second signal: the admissions officer is on holiday for the spring-break fortnight, the open morning list sits unopened, and three families who toured before the break hear nothing for 19 days. Two of them accept a place elsewhere. A CRM with a task queue would not have prevented the holiday, but a shared queue and a backup person would have made the gap visible. You can do the same in a sheet by giving the "Due now" view to a second person.

A switch-over plan that does not lose families

When you decide to move from a sheet to HubSpot free or a school-specific tool, do it in a quiet fortnight and in this order:

1. Clean first. Remove duplicates and rows with no contact consent. The CRM clean-up tutorial shows how AI can flag duplicates without changing data on its own.
2. Map your columns. Write down which sheet column becomes which CRM field and which stages become which pipeline steps. Fields that do not map are the ones to drop.
3. Import a test batch. Twenty rows. Check names, dates, stage and the consent flag on each.
4. Import the rest, freeze the sheet. Rename the old file "READ ONLY" and move it to an archive folder. Never run both for more than two or three days.
5. Rebuild the templates. Your follow-up emails go into the CRM as templates, with placeholders.
6. Review after a month. Count overdue tasks and families with no next action. Both should be near zero.

A mistake to expect: stage names. Sales tools default to "Appointment scheduled", "Qualified to buy" and similar. Rename them to your real stages before you import, or the history will confuse whoever reads it later.

Cost comparison for the first year

Using the 180-enquiry school again, here is a sensible first-year estimate. Treat it as a method rather than a quote.

• Sheet: $0 licence. Set-up: half a day to design, an hour to learn the formulas. Ongoing: about 85 minutes a week of admissions time, as above.
• HubSpot free: $0 licence. Set-up: one to two days to configure stages, forms and templates. Ongoing: perhaps 20 to 30 minutes a week saved on logging. If you outgrow the free tier and take Starter for two users at $20 a seat a month, that is $480 a year at list price.
• School-specific CRM: price on application. Ask for the first-year total including set-up, training and any per-pupil element, and compare it with the value of one enrolment.

That last comparison is the honest test. If one extra enrolment a year covers the whole cost, even a modest saving in missed follow-ups justifies it. If the tool costs more than a term's fees for one pupil, the evidence has to be strong. A tool that saves ten missed follow-ups a year is only worth what it converts, so track it for a term before you buy rather than after. For a fuller look at what AI-enabled CRMs cost a small team, see the CRM set-up cost tutorial.

What AI should not do in admissions

Keep a short list on the wall. AI does not decide who gets a place, does not score families, does not read applications or references, and does not send a message nobody has read. It drafts, it sorts, it flags. The moment it starts ranking people, you have a fairness and data-protection problem that no template solves.

One safe addition is an AI summary of your own meeting notes after a tour, written in your words and kept in the enquiry record. If you try that, summarise your notes, not recordings of the family, and read every summary before it is saved. The same discipline applies to every AI feature in a CRM: use the one that removes typing, and distrust the one that makes a judgement.

If you would like help deciding where your own admissions process sits on this scale, that is what my AI implementation consultation covers, through the contact page. Start by counting last year's enquiries and the touches each needed: the answer to "sheet or CRM" is usually already in those two numbers.

Admissions questions answered before you buy

How many enquiries justify a school CRM?

As a rough guide, a sheet copes with up to about 150 enquiries a year and one person doing admissions. Beyond that, or with two people sharing the work, missed follow-ups and duplicate contacts start to cost places. Treat the number as a prompt to test, not a rule.

Is HubSpot's free CRM suitable for a school?

It can be a good middle step for up to two users and 1,000 contacts: contacts, deals and email logging with no licence fee. It is built for sales, so you will rename fields and stages to fit admissions. Check what it stores about children, and keep sensitive notes out of it.

Can I put children's details into an AI tool?

Keep it to the minimum. Draft follow-ups from templates with placeholders, and avoid pasting names, medical or safeguarding information. If a tool processes pupil data, you need a data-processing agreement and a clear retention rule before you start.

What should happen to enquiries that never enrol?

Set a retention period, such as 12 months after the last contact, then delete or anonymise them. Tell families in your enquiry form how long you keep their details and why. A dated column in the sheet makes the clear-out a five-minute job.

Further reads

• AI Call Summaries: Log Every Phone Enquiry in Your CRM — Log phone enquiries without typing them up.
• How to Reply to Every Enquiry in Under Five Minutes With AI — Fast first replies to every enquiry.
• Tutoring Centre Admin With AI: Enquiries, Timetables, Invoices — Enquiries and timetables for a tutoring centre.
• Automate Client Onboarding With AI: Forms, Contracts, and Emails — Forms, contracts and emails once a family says yes.
• AI for a Music School: Lesson Notes and Parent Emails Worked Through — Parent emails from another small school setting.
• Answering Pupil and Parent Enquiries With AI at a Driving School — Answering pupil and parent enquiries quickly.
• What Is an AI CRM and Which Features Actually Save Time? — Which AI CRM features remove typing and which to distrust.
• AI for Private Schools: Admissions, Fees and Parent Communication — An office-side map of the private school jobs AI can do, from enquiry follow-up and offer letters to fee reminders, and where a person must decide.
• FERPA and ChatGPT: Does the Law Allow US School Staff to Use It? — What FERPA allows when staff use ChatGPT, what never to paste, and a one-page staff rule with a redaction routine for a private school.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: HubSpot main, CRM and Sales Hub pricing pages and product and services catalogue (credits), Google Sheets AI function help page, all checked November 2026.