Technology
How to implement or switch countertop shop software without losing data
A step-by-step plan for moving a stone shop onto new software: timing, data migration, testing, parallel running, cutover and training by role.
, CEO, Stonify
A software change goes well when it is run as a small project with an owner, a date and a plan. As a rule of thumb, small and mid-size stone shops take about 2 to 8 weeks from setup to full use. ERP-class systems take several months. The safe sequence is: pick one person to own it, export and back up everything, set up price lists and workflows, import the data you actually need (pricing first, active customers and open jobs next, history last), test it on real jobs, run old and new systems side by side for two to four weeks, then switch on a fixed date. Keep the old system read-only for 60 to 90 days. Train each role on the few tasks it does every day, using real jobs. In practice, rollouts fail more often because people stop using the new system than because data is lost.
Why software rollouts fail
The usual story: the owner buys a system, a vendor gives a one-hour demo to the whole team in the break room, and two weeks later two people use it while everyone else goes back to the whiteboard. The software was never set up with real prices, nobody was put in charge, and the old way was still available.
The common failure patterns are:
Bought but never configured. Price lists and workflows were never entered, so nobody can quote in it.
Partial adoption. One or two people use it; the rest work around it, so the data is never complete and nobody trusts it.
Rollback. A bad first week, no plan for problems, and the shop goes back to the old system.
Endless delay. "We'll switch after the busy season" turns into a year.
Each one is a management problem more than a software problem, and each has a fix in the plan below.
Before you start
Name one owner. Pick one person, usually the office manager or a senior estimator, not the owner who is already the bottleneck. This person configures the system, keeps the price list current, answers questions and has the authority to say "we do it in the new system now."
Write down three numbers. Measure where you are before you change anything, over 5 to 10 jobs. Good choices are quote turnaround time, template-to-install time, and remakes or callbacks per month. Without a baseline you will not know whether the change worked.
Pick your slowest month. Switch when job volume is lowest and staff have time to learn. For many U.S. shops that is January or late summer. Avoid your spring rush and the pre-holiday rush when every customer wants a kitchen by Thanksgiving. If you pick January, remember that the office is also doing year-end tax forms then: 2026 Forms W-2 and 1099-NEC (for subcontract installers, for example) are due by February 1, 2027, because January 31 falls on a Sunday. Finish them in the old system, or make sure the new one holds the whole year's payment records. If the switch includes your books, starting on the first day of a month or of your fiscal year makes opening balances easier to check.
Tell the team early. Give at least two weeks' notice, explain why the shop is changing, and say when the old system goes away.
Clean up before you move. Merge duplicate customers, fix spelling in material names, retire services nobody sells. Migrating messy data just moves the mess.
Timeline by shop size
These are rules of thumb for planning, not vendor commitments.
Shop | Typical time to full use | What takes longest |
|---|---|---|
Small (1 to 2 crews, under about 30 jobs a month) | 2 to 4 weeks | Entering price lists |
Medium (3 to 5 crews, 40 to 80 jobs a month) | 4 to 8 weeks | Price lists, integrations, training field crews |
Large (6+ crews, 100+ jobs a month, or several locations) | 8 to 12+ weeks | Data migration, integrations, multi-location setup |
ERP with full accounting | Several months | Accounting migration, inventory valuation, configuration |
A rushed switch in two weeks is possible for a small shop, but errors go up. Plan for the longer end of whatever the vendor tells you.
The step-by-step plan
Export and back up everything. Customers and contacts, open and closed quotes, job history, price lists, slab inventory, drawings and templater files, photos and attachments, and financial records. Save copies in two places, for example cloud storage and an external drive.
Map your workflow. Write down how a quote is created, how a job moves from template to install, how files reach the CNC, and how invoices go out. You will rebuild this in the new system, so it needs to be on paper first.
Configure the new system. Users and permissions by role, price lists and pricing tiers, job stages or task workflows, document templates, and integrations (accounting, payments, sales tax, texting). Entering price lists correctly is usually the biggest single task.
Run a test import on a sample. Import a slice of the data and open it: do customers have their addresses, do quotes show the right totals, are slab records complete?
Import in priority order. Pricing and catalog first, then active customers, then open jobs, then current slab inventory, and historical jobs last or not at all.
Test one real job end to end. Quote it, schedule it, send its file to the saw, invoice it. Test the templater-to-CNC file handoff on day one. It is the step where a mistake costs a slab.
Run in parallel for two to four weeks. New jobs go into the new system. Jobs already in the old system finish there. Compare quote totals and schedules weekly and fix problems as they appear.
Cut over on a fixed date. Announce it, ideally a Monday. From that day no new work enters the old system. Hand out one-page cheat sheets and have the owner and the vendor's support on call.
Keep the old system read-only for 60 to 90 days for look-ups, warranty questions and anything the migration missed. Some vendors offer a cheaper read-only or archive plan for this; ask. Cancel only after a final export is saved.
Review after 30 days. Compare your three numbers with the baseline, fix the biggest remaining problem, then turn on advanced features one at a time.
Data migration: what moves and what doesn't
Think of your data in five groups, because each can fail on its own:
Data | Usually moves cleanly? | What to check |
|---|---|---|
Customers and contacts | Yes, by spreadsheet | Duplicates, billing vs job-site addresses, builder price tiers |
Price lists and catalog | Usually, with work | Formulas and dependent charges often have to be rebuilt by hand |
Quotes and jobs | Totals yes; details vary | Whether line items come across, not only the quote total |
Drawings | Often not | Whether drawings arrive as editable geometry, as images or PDFs, or not at all |
Slab inventory and remnants | Varies | Serial numbers, sizes, costs, locations, which slabs are held for which job |
Some things rarely migrate and have to be rebuilt: custom report layouts, automation rules, custom fields (they must be mapped to fields in the new system), integration settings, user activity logs, and file links embedded in records (download the files and re-upload them).
Two rules protect you:
Export first, cancel last. Some systems delete an account's data a set time after cancellation, sometimes as little as 30 days. Never cancel before you hold a full export.
Ask exactly what comes across. "We migrate your data" can mean anything from a customer list to every drawing. Ask whether quote line items and drawings arrive as editable records, and get the answer in writing.
Checking the migration
Do not accept a migration because it "looks fine". Run a short acceptance check:
Pick a sample: a few simple jobs, a complex one, a builder account, a job with a change order.
Compare record counts by type (customers, quotes, jobs, slabs) between old and new, and explain any difference.
Open the sample records. Check prices, addresses, attached files and drawings.
Reconcile money: open balances, deposits and unpaid invoices must match to the cent.
Write down what is missing and who will fix it, and have the owner sign off.
Coming from spreadsheets and a whiteboard
Moving off spreadsheets is easier on the data side, because there is less of it. The hard part is behavior. People who have quoted from the same spreadsheet for ten years have to trust a new screen, and the whiteboard has to come off the wall. Set a date after which quotes only come from the new system, and remove the old spreadsheet from the shared drive once the parallel run ends.
Training by role
Train each group on its daily tasks with real jobs, in small groups of three to five, in short sessions. A one-hour demo for everyone rarely sticks. The times below are planning figures, not standards.
Role | What they need | Time to plan |
|---|---|---|
Office and sales | Customers, quoting, drawings, scheduling, payments | 2 to 3 hours over two sessions |
Production manager or shop foreman | Job queue, file review, machine assignments, status updates | About 90 minutes |
Templaters | Mobile schedule, uploading files and photos, field notes | 45 to 60 minutes |
Installers | Daily schedule, job details, completion photos, sign-off | About 30 minutes |
Owner and managers | Everything above plus reports, pipeline and settings | 2 to 3 hours, partly self-directed |
Methods that work in shops:
Watch one, do one, teach one. Show a task, have the person do it on a real job, then have them show the next person.
A buddy for the first two weeks. Pair each person with someone who has already learned the task.
Five-minute daily tips. One feature per day at the morning meeting beats a two-hour session.
The owner uses it too. If the boss still asks for printouts and whiteboard updates, nobody else will switch.
Expect a slower first week or two. That is normal, and it is the reason to switch in a slow month.
Handling resistance
People resist for understandable reasons. Some worry they will look slow in front of younger staff; train them one to one and start with their simplest daily task. Some think the old way worked; ask them to help configure the new system, since they know the process best. Some just wait it out; set a firm cutover date and remove access to the old tools after it.
How to tell it is working
When | Check |
|---|---|
Week 1 | Everyone has logged in; most new quotes and jobs are in the new system; people are asking questions (silence means nobody is trying) |
Weeks 2 to 4 | Quote time is falling; errors fall week over week; the old system is barely used |
Month 2 on | The old system is not used at all; your three baseline numbers are at or better than before |
What a switch costs
The subscription is often the small part. For a 12-person shop, a reasonable planning figure is roughly 40 to 60 hours of the owner's time and 60 to 90 hours of the office manager's time over six to eight weeks, plus one or two months of paying for both systems during the overlap, plus any vendor setup or migration fee. At $40 to $50 an hour including benefits, 100 to 150 staff hours comes to about $4,000 to $7,500. (For scale, the U.S. Bureau of Labor Statistics put average private-industry pay plus benefits at $46.89 an hour in June 2026.) Budget it, because a switch that runs out of time halfway is worse than no switch.
Common mistakes
Migrating all history at once instead of active data first.
Skipping the parallel run and discovering on the first busy Monday that quote totals do not match.
Not testing the templater-to-CNC file flow on a real job before cutover.
Cancelling the old system too early.
Training the office and forgetting the field crews and their app.
Choosing from a demo instead of piloting with your own jobs. See choosing fabrication software.
Letting pricing knowledge stay in one head. A switch is the moment to write the pricing rules into the system, so they do not retire with your senior estimator.
Where software helps
Stonify's import tools are built for this kind of move. Settings > Data Import > General Import takes spreadsheets for setup data such as the chart of accounts, products and services, with an example file and a blank template for each, a diagram showing which data must be imported first (for example the chart of accounts before products), and a history log of every import run. Product and service templates get one price column per pricing tier automatically. Product photos can be uploaded in bulk by file name, inventory photos have their own bulk upload, and the next project number can be set so job numbers continue from the old system. Price lists can be exported, edited in a spreadsheet and imported back, and drawing settings can be exported and imported. Roles, permissions and custom fields let the shop rebuild who-sees-what and the extra information it tracked in the old system. The feature catalog does not say whether historical quotes, open jobs or drawings from another system can be imported, so confirm that with Stonify for your specific data before planning the migration.
FAQ
Should I migrate all my old jobs? Usually not all at once. Move open jobs and active customers first. Bring history later if you need it for warranty look-ups or reporting, or keep the old system read-only and archive exports.
Can we switch in the middle of the busy season? You can, but it is harder. Staff have no time to learn, and mistakes land on real installs. If you must, run the parallel period longer and move one team at a time.
Do we need the vendor's paid training? Small shops often manage with good videos and a strong internal owner. Larger shops often find paid training worth it, because more people have to learn the system at once.
How do I get installers to use the app? Make it the only place they find the address, the job files and the balance due. Once the office stops texting those details, the app becomes the easiest option.
Sources
Related Articles
How to choose countertop fabrication software: requirements, pricing models and questions to ask
A buyer's guide for stone shops: define requirements, test vendors with your own jobs, compare pricing models and total cost, and avoid common traps.
What software does a countertop shop need? The categories and how they fit together
The software categories a stone fabrication shop uses, from CRM and quoting to CAD, nesting, scheduling, field apps and accounting, and how to connect them.
How to hire, train and keep countertop fabricators, templaters and installers
Who to hire first, what to pay in 2026, how to screen with a paid trial, how long training takes by role, and what keeps skilled stone workers from leaving.