What switching email tools really costs

Switching email tools costs a small firm a few thousand dollars. Almost none of it arrives as an invoice. It’s hours: the hours you spend deciding, and the hours everyone works a little slower while the new app is still new. Price those two honestly and you’ve got your number.

• The biggest single line is the deciding. Comparing three products, sitting through demos and talking it over internally can cost the owner more than connecting, learning and settling in put together.
• The second biggest is per head, and it multiplies. Everyone’s slower for a couple of weeks. A firm of eight pays that eight times over, and a firm of one pays it once.
• The money you can’t get back is the remainder of the term on whatever you’re replacing. That figure comes from the terms of the tool you’re leaving, and vendors differ enormously on it.
• There’s no data migration line, because nothing moves. That’s the one large cost this kind of switch doesn’t have, and it’s the cost most people budget for.
• A switch you abandon halfway costs nearly the full amount and buys nothing. That’s the real argument for deciding once, carefully, instead of trying four things in a row.

What a switching cost actually is

There’s a clean definition in the economics literature. It’s worth borrowing, because it tells you exactly where to look.

Joseph Farrell and Paul Klemperer, in the switching costs section of their chapter for the Handbook of Industrial Organization, put it this way: “A consumer faces a switching cost between sellers when an investment specific to his current seller must be duplicated for a new seller” (Coordination and Lock-In, section 2.1). The investment, they say, “might be in equipment, in setting up a relationship, in learning how to use a product.” They also draw a line that matters here. Some switching costs are learning costs, paid once and never again. Others are transactional, and they get charged again if you ever switch back.

So the whole question is which of your investments are specific to the seller you’d be leaving. Write two lists.

Specific to your provider. Your address. Fifteen years of archived mail. Your contacts. Your domain and its DNS records. Your calendar. Your server-side filters. Your aliases. Your shared info@ mailbox. This list is enormous, and it holds almost everything you’ve ever put into email.

Specific to the app you read mail in. Your signature. The layout your hands know. Whatever rules you built inside that particular app rather than on the account. The preferences you taught it. Not much, once you write it down.

Changing email clients touches only the second list. The mailbox stays where it is, and the new app signs in to it. Changing providers touches the first list, which is why that job takes a weekend and a plan. Why there is no migration when you switch email apps works through the difference properly, and the checklist for the day itself is how to switch to an AI email client.

That’s the good news, and it’s also why the rest of this page exists. When people hear there’s no migration, they conclude the cost is zero. It isn’t zero. It’s moved into a column nobody itemizes.

The seven things a switch actually costs

Here’s the whole bill for a firm changing the app it reads mail in. The sizes are rough, for a six-person practice. Fill in your own hours. What matters is which lines are large.

Four of those seven are small enough to ignore. Connecting is a secure sign-in and a wait while the app reads back through mail that’s already there. Setting up what doesn’t travel means a signature pasted in and a sending address checked. You turn the old app’s notifications off before the new app’s go on. Teaching the app your habits is real time. It’s time spent getting the thing you bought, and it returns more than it takes.

Three lines decide your number: the deciding, the working slower, and the remainder on the old subscription. Each gets its own section below. Each is mispriced in a different direction.

Look at what the bill leaves out. Nobody exports mail. Nobody imports it. No consultant, no downtime, and nothing to tell clients. In an ordinary software switch those five lines are most of the budget. Here they’re all zero, for the reason the first section gives: none of them is specific to the app.

What an hour of your firm’s time is worth

Every number on this page is hours until you price it. Firms price these hours badly, and always in the same direction. Too low.

The floor is the wage. For the year ending May 2025 the Bureau of Labor Statistics puts the median annual wage for accountants and auditors at $83,680, which it also states as $40.23 an hour (Occupational Outlook Handbook, checked September 7, 2026). Bookkeeping, accounting, and auditing clerks come in at $50,670, or $24.36 an hour (Occupational Outlook Handbook, checked September 7, 2026). If your practice is consulting rather than tax, the comparable figure for management analysts is $101,860, or $48.97 an hour (Occupational Outlook Handbook, checked September 7, 2026).

Those are national medians for wages, so treat them as an anchor rather than as your payroll. Two adjustments make them honest here.

First, an hour costs the firm more than the wage. Payroll taxes, benefits, software and rent all ride on top. The usual working assumption is that a fully loaded hour runs well above the wage that produced it.

Second, and this one is larger. For anyone who bills, an hour spent on this costs you the hour that didn’t get billed. A partner reading comparison pages on a Tuesday afternoon in March costs the firm a good deal more than $40. Use your own realization rate. You’ll get a number two to four times the wage line, and that’s the right one.

The owner’s hour is the expensive one, and the owner spends the most of them. That’s the shape of the whole exercise. It’s also why the next section is about the part that never gets written down.

The line nobody puts on the list

Ask a firm what a switch cost and you’ll hear about the week they connected. Ask what happened in the two months before that and you get a shrug. None of it looked like work on the project.

It was the work, though, and here’s what it’s made of. Reading around the category. Two or three vendor calls, each an hour, with a half hour of scheduling and follow-up attached. Watching demos. Building a comparison in a spreadsheet nobody opens again. Asking two people whose judgment you trust. Raising it at a partners’ meeting twice, because the first time nobody had an opinion. Sitting on it for three weeks. Coming back to re-read the same three pages, because you’ve forgotten what told them apart.

For most owners that comes to somewhere between six and fifteen hours. Every one is an owner hour, at the highest rate in the firm. In a six-person practice it’s routinely the biggest number in the whole switch, larger than the entire staff’s slower two weeks put together.

Two things follow, and you can act on both.

Shorten the deciding, and the switch gets cheap. Almost every article about switching software is about executing well, and execution is the cheap part here. To make your number smaller, set a date by which you’ll have decided. Write down in advance what would make the answer yes. Then stop reading once you have those answers, even if you don’t feel certain. Certainty isn’t on offer, and waiting for it is billed by the hour.

A switch abandoned after the trial still costs you. You pay the whole deciding line. You pay the connecting. You pay a week or two of everyone being slightly slower, and you get nothing else. Go again in eight months and you pay the deciding line a second time, because you won’t remember what you concluded. So make the trial decisive, with a verdict at the end of it, and the guide for that is trying a new inbox without the risk.

The month you pay for both

One line here arrives as an actual dollar figure, and that’s the overlap. The exit terms of the tool you’re replacing set it. Read those terms before you start, because they differ by more than the products do.

Two live examples, both from the vendors’ own pages, both checked September 7, 2026.

Calendly’s Customer Terms and Conditions state plainly that “You understand that all Fees are non-cancellable and non-refundable and must be paid without offset or deduction of any kind,” and that “The Services shall automatically renew for successive Renewal Subscription Terms for a term length equal to the previous subscription period” (Customer Terms and Conditions). If you’re seven months into an annual term when you decide, five months of that term are already spent money.

SaneBox goes the other way. Its Terms of Service say the company will try to fix a problem first, and “If we can’t, we will cheerfully refund the amount within 30 days of purchase” (Terms of Service).

Same kind of purchase, opposite consequence for the timing of your decision. The practical job takes five minutes. Before you start evaluating, open the terms of every tool the new one might replace. Write down two things: the renewal date, and whether anything is refundable. That tells you what your overlap costs, and often when to start.

Two smaller notes on the same line. You keep paying your provider throughout, because the mailbox comes from Google or Microsoft and the client is an app pointed at it. You also plan to run both apps for a couple of weeks, and that’s money well spent, since it’s what lets you back out cheaply. The hidden cost is the annual term you forgot you were on.

The wider version of this exercise covers the order to cancel things in, and what breaks when a subscription stops. That’s how to find and cut overlapping subscriptions. Here you only need the two dates.

Why the number grows while you wait

Farrell and Klemperer’s definition has an uncomfortable implication if your plan is to think about this next year.

The switching cost is the investment specific to the current seller, and that investment grows every month you stay. Every month adds a little more of it. Another saved search. Another rule. Another person who joined and learned the current way. Another booking link handed to another client. Another habit that only makes sense inside this particular app. Nobody notices, because it arrives in pieces too small to see.

So the cheapest day to switch was some time ago, and the second cheapest is today. Waiting costs you a small monthly payment into the price of the thing you’ve already decided you’ll probably do.

One more point explains a lot of vendor behavior. Your switching cost is an asset on your current supplier’s balance sheet. It’s why a product you’ve used for four years can raise its price, retire the tier you liked, or stop improving, and keep you anyway. That’s what the economics predicts, and the defense is to know your own number rather than to feel it. A firm that’s priced its switching cost at $4,000 makes a different decision about a $60 a month increase than a firm that’s priced it at infinity.

Here’s the consoling half of the same idea. Most of what you’d pay is learning rather than transactional, so you don’t pay it again if you go back. What you learned about email stays with you, and the mailbox never moved. Going back is cheaper than going forward was.

Where the estimate goes wrong, both ways

Firms get this number wrong in two directions at once. That’s why the guess and the reality so rarely look alike.

Overestimated, every time:

• The data. No export. No import. No reconciliation, and no weekend. People budget for a migration out of memory, because the last email project they lived through was one.
• The disruption to clients. It’s zero. Your address stays the same, your threads stay threaded, and there’s nothing to announce, which is the subject of why you don’t need a new email address.
• Their own learning curve. Owners think they’ll personally take longer than they do. Reading a sorted list comes easily.
• The setup. A signature and a notification toggle. Firms brace for an afternoon and spend eleven minutes.

Underestimated, every time:

• The deciding. As above. It hides, because it’s spread across eight weeks in twenty-minute pieces.
• Everybody else’s learning curve. The owner chose this and is motivated. The bookkeeper didn’t choose it and isn’t. That gap is a real cost with a real remedy, which is what bringing your team along on a new tool is about.
• The multiplication. Every per-person line multiplies by heads, and so does the subscription. A firm of twelve pays twice on every line at once.
• The tail on the old tool. Somebody has to close it properly. Export what was in it. Take the seats off. Remove the links clients hold. Remember the automation nobody thought about. It’s a real afternoon, and it lands six weeks after everyone stopped caring.
• The second attempt. Abandoned switches are common, and they cost close to the full amount.

A firm of six, totaled

Put it together for a practice with an owner, three accountants, a bookkeeper and an administrator. Use your own rates. The shape is the useful part.

The owner. Twelve hours deciding, then eight hours across connecting, learning and closing out the old tool. Twenty hours. Priced at an hour that would otherwise bill, this is comfortably the largest line. At $150 an hour it’s $3,000, and in most practices the true figure is higher.

Everyone else. Call it eight hours each of slower work, setup and teaching, spread over the first month. Forty hours. At the BLS medians above, three accountants at $40.23 and two staff at around $24.36 come to roughly $1,350. At what those hours would have billed, two or three times that.

The overlap. The unexpired remainder on whatever the new tool replaces. Say a scheduling seat and a drafting add-on with four months left on annual terms: a few hundred dollars.

Total. Somewhere between $4,500 and $5,000, with the owner’s hours at a billed rate and everyone else’s at wages. Past $7,000 if you value the staff hours at what they would have billed too. It’s one time, spread across roughly six weeks, and about two thirds of it belongs to one person.

Now put it next to the recurring cost. Full AI email clients sit at roughly $25 to $45 a seat a month. Those figures come from what an AI email client costs, which reads them off the vendors’ own pages with dates, and counts the seats firms forget. Six seats lands somewhere near $1,800 to $3,240 a year.

So for a firm this size, switching costs about one and a half to three times the first year’s subscription. That’s the sentence to carry away, because it changes the question. You’re deciding whether the software is worth roughly $200 a month plus one large, one-time payment of the firm’s attention. If the answer is yes, it’s yes by a distance, and sooner is cheaper. If it’s marginal on the subscription alone, the switching cost settles it against you, and you leave it alone.

Two adjustments for a different-shaped firm. Solo, the multiplication disappears and the deciding is the whole cost, so a solo practitioner’s real risk is spending eleven hours choosing between two products that are both fine. Above about fifteen people the per-head lines take over. The team’s adoption becomes the project, and the arithmetic starts to look like ordinary change management.

What makes the number smaller

Five levers, in order of how much they move the number.

Decide on a deadline. Write down the three things that would make the answer yes, give yourself two weeks, and stop. This one is worth more than the other four combined, because it works on the biggest line.

Change one thing. Most of what a switch costs is the confusion. Keep your filters. Keep your old app installed. Keep everything else exactly as it was, and every strange thing that happens in the first two weeks has one possible cause. Firms that clean house at the same time double their diagnosis time, then blame the software for the rest.

Pick the week deliberately. Choose a slow one. Stay clear of the first week of April, the week before an October 15 deadline, and any week when someone’s on vacation. A slow week in June costs a fraction of a busy one, and the choice of week is free.

Send one person first. One person a week ahead of everyone else finds the four surprises. Then five people never spend an hour each on them. It also gives you the only trustworthy evidence about whether the rest of the firm should go at all.

Close the old thing on a date. Put the shutdown in the calendar the day you start, with a name against it. The tail is what turns a four-week project into a four-month one, and a date is what prevents it.

Most of the anxiety goes somewhere else: the technical setup, the choice of afternoon, and how much of your mail history the new app reads back. Those three are already small.

The cost of not switching

The switching cost means something only next to what you’re weighing it against, and that comparison has to be honest in both directions.

What email already costs a firm has been measured, and the numbers are large. What the logged studies found, and how to count your own week rather than borrow a statistic, is in how much time firms lose to email, measured. Email never gets its own account code, so those hours land oddly on the books, and that’s the hidden cost center your firm calls an inbox. Against a standing annual cost of that size, a one-time four or five thousand dollars pays back quickly, and how quickly it starts is the subject of getting value in the first week.

And here’s the case where the answer is no, said once and plainly. Your inbox is manageable. Nothing important has been missed. You get through the day’s mail without dread. Then there’s nothing for a new tool to recover, the switching cost is the whole story, and you keep your money. The cost question becomes a value question when you have the problem. Plenty of firms don’t.

What a switch to Point costs

Point is an AI email client, so the arithmetic above is the arithmetic that applies. Here’s which lines Point makes smaller, and which ones stay where they are.

The migration lines go away, and they go away by design. Your address, your history and your contacts stay where they are, because Point reads that mailbox where it sits. You connect the Gmail or Microsoft 365 account the firm already has, through a secure sign-in, with no forwarding to set up. Nobody gets an announcement, and your provider subscription carries on unchanged underneath.

The per-person lines get smaller, for one reason. Point writes its decisions back into the mailbox you already have. What Point triages arrives in real Gmail labels, or Outlook categories, so the sorting shows up in the old app too and running both apps for two weeks actually works. Archive a thread in Point and the same thread leaves the inbox on the provider’s side, filed and still there. That’s what makes the overlap a real safety net. You teach Point in a sentence, and what Point picks up is kept as plain lines you can read and change, which is why that line is an hour and not a training day.

The two large lines stay. You still have to decide, and no page can spend those hours for you. Everyone still works a little slower for a couple of weeks. Point is a per-seat subscription, priced by the seat during the private beta with no published list, so the recurring line grows when you hire and shrinks when someone leaves.

Two things decide what a bad first two weeks would cost you, and nobody prices this part of a switching decision. The first is how much Point does on its own while you’re still forming a view. Every category of work carries its own position, and all of them arrive set to review, which means Point prepares the thing and waits for you. A rung below is suggest-only. A rung above is fully handled, where Point gets on with that work without checking, so you move to that position on purpose rather than on the first afternoon. The second is reversibility. What was done is listed in plain words with times beside it, and most rows go back from the list itself. The exception is the one nothing in this category escapes, which is mail that has already landed on the recipient’s own server.

If you leave, the mailbox is the mailbox you always had, and what stays behind and what goes is set out in if you ever want to leave. Everything Point does is the full list to hold against your own week before you spend an hour on the comparison.

Common questions

How much does it cost to switch email tools?

For a six-person firm, roughly $4,500 to $5,000, and past $7,000 once you value the billable hours at what they would have billed. Almost all of it is time rather than money: the owner’s hours spent deciding, plus a few hours each of everyone working slower for two weeks. The one real cash line is the unexpired remainder on whatever subscription you’re replacing.

Is there a migration fee, or a setup cost?

There shouldn’t be, and a fee is a reason to ask hard questions. Changing the app you read mail in takes no export, no import and no server work, so a fee has nothing to pay for. An implementation charge on a product you adopt by signing in charges you for the one line this kind of switch doesn’t have.

Which part of the cost is the biggest?

The deciding, in almost every practice of this size. Six to fifteen owner hours, spread across two months of reading, calls, demos and internal conversation, at the highest hourly rate in the practice. It’s the biggest line and the least visible one, because none of it looks like work on the project while it’s happening.

How long is the firm slower for, and by how much?

Nobody stops working. Budget four to ten hours per person, spread across two weeks rather than lost in a block. Everyone’s a bit slower at finding things while the shape of the new app settles, and then the speed comes back. The owner is usually back to normal speed first, which is exactly why owners underestimate this line for everyone else.

Do we have to pay for two tools at the same time?

For a few weeks, and you should want to. Running the old app alongside the new one is what makes backing out cheap. The cost that matters is the unexpired remainder of an annual term on a tool you’re dropping, which is why the renewal dates are worth checking before you start evaluating rather than after you’ve chosen.

What does it cost if we try it and decide against it?

Nearly the full amount, minus the closing-out. You pay the deciding, the connecting and a week or two of everyone being slower, and you get an answer back. So trying is fine. Make the trial decisive, with the questions written down in advance, so the hours produce a verdict rather than a feeling.

Does the cost go up the longer we put it off?

Yes, slowly and quietly. Every month adds a little more of the investment you’d have to duplicate. Another rule. Another habit. Another new hire who learned the current way. Another link handed to another client. Waiting has a price, which is worth knowing if the plan is to revisit this next year.

Does it cost our clients anything, or cause them any disruption?

No. Your address stays the same, and threads stay threaded. The mail never leaves the account it’s already in, so there’s no window where it could go astray. There’s nothing to announce, and nothing here your clients could detect.

How does the switching cost compare to the subscription?

For a six-person firm it runs about one and a half to three times the first year’s subscription, which makes year one roughly double year two. Hold that number in your head when you approve the monthly figure. It’s the reason a marginal case should be a no. A clear yes pays it back quickly; a maybe doesn’t.

Is it cheaper to switch at a particular time of year?

Meaningfully, yes, and it’s the cheapest lever you have. A slow week costs a fraction of a busy one for the identical work, because the hours everyone spends being slower are worth less when there’s slack in them. For a tax practice that means avoiding April and the run-up to October 15. Choose the week the way you’d choose the day for anything else disruptive.

What it comes to

• The bill for changing email tools is hours, and two lines decide it: the owner’s time spent deciding, and everyone else’s two weeks of working slower. There’s no migration line, because nothing moves.
• Price the hours at what they would otherwise have billed, rather than at the wage. For a six-person practice the total lands around $4,500 to $5,000, one time, and roughly one and a half to three times the first year’s subscription.
• The one cash figure is the unexpired remainder on what you’re replacing, and that vendor’s terms set it rather than yours. Read the renewal dates before you start looking.
• Shorten the deciding, change one thing at a time, pick a slow week, send one person first, and put the shutdown of the old tool in the calendar on day one. Those five move the number more than anything technical does.

If the question underneath the cost is whether you can try this without exposure, that’s trying a new inbox without the risk. If it’s what happens to your address, you don’t need a new one. If it’s whether the rest of the firm will come with you, bringing your team along is where the per-head lines get cheaper. And Point is what the number buys: a morning that opens on mail already weighed and put in order.