---
title: How to move your business off Excel without stopping work | Efidis
description: When Excel is the right tool, what it costs to run on it, 12 signs it's time to move and a 9-step plan to migrate one area without shutting anything down.
url: https://efidis.co/en/blog/como-salir-de-excel
lang: en
---

# How to move your business off Excel without stopping work

*Operations · Efidis team · Updated November 4, 2026*

**In short:** Excel isn't the problem; using it as your operation's shared system is. In a study of 25 real working spreadsheets from five organizations, 16 had confirmed errors, and the largest was worth about $100 million (Powell, Baker and Lawson, 2009). To move off it without stopping the business: list every spreadsheet you run on, start with the process that hurts most, clean up your master data, migrate only open balances, run the system in parallel for one full close cycle and retire the spreadsheet once the numbers match. For one area, plan on 10 to 12 weeks.

## Excel isn't the problem

Before we talk about migrating, let's get one thing straight: Excel is a great tool. Banks, universities and nearly every company you know use it, and for good reason. If someone tells you to get it out of your business, be skeptical.

The trouble starts when a spreadsheet stops being one person's tool and becomes the system your operation runs on: the place where orders get taken, stock gets tracked or receivables get managed. We call that the shadow system. Nobody designed it as a system and nobody approved it as one, but the business depends on it every day.

Excel is fine, and you should stick with it, when most of this is true:

- One person maintains it and uses it. Whoever built the sheet understands it, and that's enough.
- It's analysis, not operations: forecasting sales, comparing scenarios, building a budget, checking margins.
- You're testing a new process. A spreadsheet is the cheapest way to figure out what data you need before you pay for a system.
- Volume is low and steady: a few dozen orders a month, one shipping location, a short product list.
- Mistakes show up quickly and are cheap to fix.

If that's you, don't migrate. Clean up the spreadsheet (there are ideas for that below) and spend the money on something else.

You can tell you've outgrown it by three things. First, several people write in the same file, or in copies of it. Second, data moves from one sheet to another by hand. Someone copies today's orders into the shipping sheet, then copies shipments into the billing sheet. Third, the business makes real decisions off those numbers: who gets shipped, what gets bought, how much a sales rep gets paid.

At that point Excel isn't failing because it's bad. It's failing because you're asking it to do something it wasn't built for: be a shared database with permissions, an audit trail and rules nobody can skip.

|  | Excel works well | Excel falls short |
|---|---|---|
| Who edits | One or two people who coordinate | Five, ten or more people, sometimes at the same time |
| Type of work | Analysis, budgets, scenarios | Orders, stock, receivables, shipping |
| Rules | The person who built the sheet remembers them | They have to hold every time: prices, credit limits, approvals |
| Permissions | Everyone sees everything and that's fine | A sales rep shouldn't see margins or other reps' accounts |
| Access | From the office computer | From the road, the warehouse or a phone |

## What it costs to keep going this way

The costs of a shadow system almost never show up on your P&L: errors nobody measures, hours nobody counts and a dependency nobody admits to. Here's what the research actually says, without overselling it.

### Spreadsheet errors are normal

Raymond Panko, a professor at the University of Hawaii, has been collecting spreadsheet audits by other researchers since the 1990s. His bottom line is that spreadsheet errors are common, and the sensible move is to assume they're there rather than assume a sheet is fine because nobody has complained (Panko, 2005). In field audits of real spreadsheets since 1995, 94% of the 88 spreadsheets reviewed across seven studies had errors.

The study most useful to a small business is a different one, because it looks at real working files instead of lab exercises. Stephen Powell, Kenneth Baker and Barry Lawson reviewed 25 operational spreadsheets from five organizations, the kind used to make decisions every day. They found 117 confirmed errors in 16 of them. Only 9 were clean. The costliest error was worth about $100 million (Powell, Baker and Lawson, 2009).

Think about what that means: nearly two out of three spreadsheets those organizations relied on had at least one confirmed error. And they were in use.

JPMorgan, 2012. The bank lost about $6.2 billion on a derivatives position known as the "London Whale" trade. Its own internal report says the model used to measure the risk ran on Excel spreadsheets that were filled in by copying and pasting data by hand, and that one formula divided by the sum instead of the average. The nuance matters: the report blames the loss on the trading strategy, not the spreadsheet. What the error did was silence an alarm that should have gone off (JPMorgan Chase, 2013).

Reinhart and Rogoff, 2013. Two Harvard economists had published a widely cited study on public debt and growth. Three researchers at the University of Massachusetts tried to reproduce it using the original spreadsheet and found that an average didn't cover the full range of rows: it left out 5 of the 20 countries (Australia, Austria, Belgium, Canada and Denmark). With all the corrections they proposed, average growth for countries with debt above 90% of GDP went from −0.1% to +2.2%. Careful, though: that swing combines several corrections, not just the Excel error (Herndon, Ash and Pollin, 2013).

Genetics, 2016. Excel turns some gene names into dates: SEPT2, for example, becomes "2-Sep." A team of researchers reviewed scientific papers that published gene lists in Excel and found the problem in 704 of 3,597, about one in five (Ziemann, Eren and El-Osta, 2016). A review published five years later found it in 30.9% of papers (Abeysooriya et al., 2021). Highly trained people, peer review, and auto-formatting still won.

What does any of this have to do with a distributor or a small manufacturer? Same mechanism, smaller scale. A price list every sales rep keeps in their own version. A commission formula someone dragged one row short. A product code Excel turned into a number, dropping the leading zero. An order pasted twice into the shipping sheet. None of it makes the news, but each one ends in a wrong shipment, a corrected invoice or an argument with a customer.

### The hours lost to copy and paste

The second cost is time. In a Smartsheet survey of office workers, over 40% said they spend at least a quarter of their week on manual, repetitive tasks. The biggest time sinks were email, data collection and data entry (Smartsheet, 2017).

The McKinsey Global Institute got to a similar place from another angle. Looking at which parts of work can be automated with technology that already exists, it found that collecting data (64% of the time) and processing it (69%) have the most technically automatable time of any activity after predictable physical work (McKinsey Global Institute, 2017).

Run the numbers for your own business. If your receivables person spends half of every Monday matching the invoice sheet against the bank statement, that's about 200 hours a year. If three sales reps text the warehouse five times a day to ask what's in stock, it adds up fast. You don't need a study to know how many hours it is. Ask your team. The [formula for what manual data entry costs](https://efidis.co/en/blog/cuanto-cuesta-copiar-datos-a-mano) turns those hours into money.

### The business depends on one person

The third cost is the one you don't see until it hurts. In almost every company that runs on spreadsheets, there's someone who understands the big file: where the formulas are, which macro runs on Fridays, why nobody touches column K. That person is valuable, and also a risk. If they get sick, go on vacation or quit, part of the operation goes dark. A system doesn't replace people who know the business, but it puts the rules in the system instead of in someone's head.

## Self-check: 12 signs it's time

Check the ones that apply to your business today, honestly: the goal is to know where you stand.

- [ ] More than three people edit the same file, or copies of it.
- [ ] You have files named something like "inventory_final_v7_USE_THIS" and someone has to ask which one is current.
- [ ] Every week someone copies data from one sheet to another by hand: orders to shipping, shipping to billing.
- [ ] The monthly report takes more than two days because several files have to be pulled together.
- [ ] Sales reps call or text the warehouse to ask what's in stock.
- [ ] You can't give someone access to part of the information without showing them the whole file.
- [ ] Only one person understands the formulas or macros in the main file.
- [ ] A spreadsheet error has already cost you money at least once: a double shipment, an old price, a commission paid wrong.
- [ ] When a number changes, you can't tell who changed it or when.
- [ ] The file is slow to open, gets corrupted, or had to be split into several by year or month.
- [ ] Field or warehouse staff can't look anything up or record anything without calling the office.
- [ ] To know how the month is going, you have to ask several people for numbers and wait.

> If you checked 3 or more, it's time. You don't have to migrate everything: pick the process that hurts most and start there. If you checked fewer than 3, you'll probably get more out of cleaning up your spreadsheets than replacing them: one master sheet per type of data, data validation on cells, view-only access for anyone who doesn't edit, and a dated backup every week.

## 9 steps to move off Excel

This is the order that works. The first four steps are diagnosis and don't require buying anything; the other five are the migration itself. Skipping the first four is the fastest way to end up with a new system that repeats the spreadsheet's problems.

1. **Take stock of your spreadsheets.** Every file the operation runs on, who updates it and what happens if it breaks.
2. **Map the flow.** Follow an order to payment and mark where data gets copied by hand.
3. **Decide where each process goes.** Clean up the sheet, off-the-shelf software or a custom system.
4. **Pick where to start.** The one that hurts most and depends least on others. Just one.
5. **Clean up your master data.** No duplicates, and one owner per type of data.
6. **Decide what to migrate and what to leave.** Open balances, yes; history, almost never. With a cutover date.
7. **Run both in parallel.** One full close cycle, comparing numbers every week.
8. **Train by role.** What each person does on day one, not the whole system.
9. **Retire the spreadsheet.** With written criteria and a sign-off.

### Step 1. Take stock of your spreadsheets

Before you think about systems, find out what you have. This almost always turns up surprises: files nobody knew existed that someone has been updating every week for years.

Do it in one afternoon with the head of each area. Ask each of them to open their folder and name the files they use to get work done, not the ones they glance at now and then. Log each file in a table like this:

| File | Process | Owner | Who updates it | Where the data comes from | Who it feeds | Frequency | Criticality (1–3) |
|---|---|---|---|---|---|---|---|
| Orders 2026 | Order entry | Sales manager | 4 sales reps | Texts and phone calls | Warehouse | Daily | 3 |
| Warehouse stock | Inventory | Warehouse manager | Warehouse assistant | Packing slips and counts | Sales and purchasing | Daily | 3 |
| Current prices | Price list | Management | Management | Suppliers | Sales reps | When prices change | 3 |
| Receivables | Collections | AR clerk | AR clerk | Invoices and bank statements | Management | Weekly | 2 |
| Commissions | Sales rep payouts | Accounting | Accounting | Orders 2026 | Payroll | Monthly | 2 |

Criticality is 3 if the operation stops when the file breaks or has an error, 2 if you lose money or time but can keep going, and 1 if it's reference only.

Two columns tell you more than the rest. "Where the data comes from" shows where the same thing gets typed twice. "Who it feeds" shows the dependencies: if Commissions pulls from Orders, you can't migrate one without thinking about the other.

### Step 2. Map the flow, not the columns

The list tells you which files exist. The flow tells you how information moves between them, and that's where the real cost is.

Take one process end to end. For a distributor, for example: order, stock check, shipment, invoice and payment. For each step, answer four questions:

- Who does it, and with what? The spreadsheet, the phone, text messages, a notebook.
- What data does it receive, and from where?
- What data does it hand off, and to whom?
- Where does someone copy something by hand?

Do it on a big sheet of paper or a whiteboard, with arrows. You don't need any software. What matters is marking in red every point where a person moves data from one place to another. Every red mark is time you're losing today and a spot where errors get in.

You'll also find rules nobody ever wrote down: "don't ship to this customer if they're more than 60 days past due," "this product only sells by the full case," "orders that come in after 3 ship the next day." Write them all down. Those are the rules the new system has to follow, and if you don't capture them now, they'll surface on go-live day.

One piece of advice: map the flow the way it is, not the way it should be. If the GM approves discounts over text, that goes on the map.

### Step 3. Decide where each process goes

Moving off Excel doesn't mean putting everything into one system. Each process on your list has four possible destinations:

| Destination | When it fits | Examples |
|---|---|---|
| Improve the spreadsheet | One or two people, low volume, more analysis than operations | Annual budget, forecasts, margin analysis |
| No-code tool (forms, online databases) | A simple, self-contained process without many rules or links to others | Visit logs, internal requests, equipment tracking |
| Off-the-shelf software | Processes every company runs the same way, or that follow outside regulations | Accounting, payroll, e-invoicing |
| Custom system | Processes that set you apart, with your own rules, or that are spread across several sheets feeding each other | Orders with customer-specific pricing, multi-warehouse inventory with reservations, routes and field collections |

The question that helps most: is the way you run this process the same as at any other company, or is it part of why customers buy from you? Accounting works the same everywhere; buy off-the-shelf software and adapt to it. The way your reps take orders in the field, with your price lists and credit limits, maybe not. To decide with a score, use the [scorecard for custom vs. off-the-shelf software](https://efidis.co/en/blog/software-a-la-medida-o-estandar).

For most small businesses the answer is a mix: accounting in off-the-shelf software and operations (sales, inventory, shipping) in a custom system connected to it. What you want to avoid is ending up with five apps that don't talk to each other either.

### Step 4. Pick where to start

With the list and the flow in front of you, pick one process to start with. Just one. The test is simple: the one that hurts most and depends least on the others.

"Hurts most" comes from what you already wrote down: criticality 3, lots of red marks on the map, errors that already cost money. "Depends least" means you can migrate it without moving five other things at the same time. If inventory feeds orders, purchasing and commissions, it may make sense to start with orders and include inventory, instead of trying to move all four at once.

Why not everything at once? Because big software projects run over more. A McKinsey and University of Oxford study of more than 5,400 IT projects found that large ones, over $15 million, ran 45% over budget and 7% over schedule on average, and delivered 56% less value than planned (Bloch, Blumberg and Laartz, 2012). Another analysis of 1,471 projects found that one in six was a "black swan": in those cases the average cost overrun was 200% and schedules ran almost 70% long (Flyvbjerg and Budzier, 2011).

Your project isn't $15 million, but the mechanism holds at any size: the more you try to move at once, the more you discover late and the more the scope grows. In Panorama Consulting's 2026 report on ERP projects, the two most cited causes of budget overruns were adding technology that wasn't planned (54.9%) and expanding scope (51.0%) (Panorama Consulting, 2026).

### Step 5. Clean up your master data

Master data is what everything else uses: customers, products, prices, warehouses, sales reps, payment terms. If it's messy, the new system starts messy, and within the first week your team decides "the system is broken."

Cleanup almost always takes longer than loading. Start as soon as you've picked the process, without waiting for the system to be ready. Here's the list to work through:

- [ ] Customers: one record per customer, with tax ID, shipping address, contact, assigned rep and payment terms.
- [ ] Products: one unique code per product, with no two codes for the same item and no code shared by two items.
- [ ] Units: each product with its selling unit and, where it applies, its conversion (case of 12, pack of 6).
- [ ] Prices: only current price lists, with the date they take effect and which customers they apply to.
- [ ] Warehouses: each one with a fixed name and code, not "the new warehouse" and "the one out back."
- [ ] Sales reps: who's who and which customers or territories each one covers.
- [ ] Payment terms and credit limits: per customer and in writing, not from memory.
- [ ] One owner for each type of data: the person who decides what's correct and who, after go-live, is the only one who changes it.

To find duplicates in a sheet before you migrate, three simple tricks:

- Count repeats. In a new column, enter =COUNTIF($B:$B,B2), where B is the tax ID column. Anything above 1 is a duplicate: filter and review.
- Normalize before you compare. "Lopez Hardware," "LOPEZ HARDWARE" and "Lopez Hardware " are the same customer to you, but not to Excel. =UPPER(TRIM(B2)) strips extra spaces and standardizes capitalization; check abbreviations like "Inc." and "LLC" separately with find and replace.
- Protect codes with leading zeros. If code 00123 shows up as 123, Excel turned it into a number. Format the column as text before you paste the data in.

Don't clean up alone. Sit down with whoever knows each piece of data: your sales manager knows which of those three "Lopez" entries are the same customer, and you might not.

### Step 6. Decide what to migrate and what to leave

The temptation is to bring everything: five years of orders, every invoice, every stock movement. It's almost never worth it. Migrating history takes time, forces you to clean old data nobody will use and pushes back go-live.

The rule of thumb: migrate what's open and keep what's closed for reference.

| Data | Migrate it? | How |
|---|---|---|
| Customers, products, prices, warehouses | Yes | Already cleaned up (step 5) |
| Stock by warehouse | Yes, as opening balances | Backed by a physical count on the cutover date |
| Receivables | Yes, invoice by invoice | Everything still unpaid, with its due date |
| Open orders | Yes | Anything not yet shipped or invoiced |
| Sales history | It depends | If you need it for reports, summarized by month, customer and product |
| Closed orders, invoices and stock movements | No | They stay in the old files, view-only |

The key to this step is the cutover date: the day from which everything is recorded in the new system. Pick a slow day, ideally at the start of a month, and announce it well ahead. That day you count inventory, close out receivables and load whatever is still open as opening balances.

### Step 7. Run both in parallel

For a while, the operation gets recorded in the new system and in the spreadsheet. It's extra work, and it's the only way to know the system is right before you depend on it.

How long? One full close cycle: if you close monthly, one month. Less than that leaves processes untested, like month-end billing, commission payouts or the receivables cutoff. More than that wears your team out, they start filling in only one of the two, and the parallel run stops being useful.

What to compare:

- Stock: every week, a physical count of a sample of products against what the system says.
- Receivables: at the end of each week, the balance per customer in the system against the spreadsheet.
- Sales: every day, the total per rep in the system against the spreadsheet.
- Orders: every order in the spreadsheet is in the system, and vice versa.

When a discrepancy shows up, figure out where it came from before you fix the number. It's almost always one of three things: master data loaded wrong, a business rule nobody told the system about, or an error that was already in the spreadsheet and nobody had caught. That last one happens more than people think, and it's good news.

Name one person who signs off every week that the numbers match, or explains why they don't. Without that person, the parallel run turns into two sets of records nobody compares.

### Step 8. Train by role

The most common kind of training is the worst: a two-hour session for everyone where someone demos the whole system. The next day nobody remembers anything, because nobody has to use the whole system.

Train by role. For each one, define what that person has to be able to do on day one, and practice only that, with real data:

| Role | What they need to do on day one |
|---|---|
| Sales rep | Look up a customer, check stock, enter an order with the right price list and check an order's status |
| Warehouse | See orders waiting to ship, record a shipment, receive a purchase and make an adjustment with a reason |
| Accounts receivable | See receivables by customer and by aging, record a payment and put a customer on or off hold |
| Management | See today's sales, stock and receivables without asking anyone for numbers |
| Admin | Create users, change prices and fix master data, with a history of every change |

Three details make the difference. Train with your own company's data, not a generic demo. Do it close to go-live, not three weeks out. And in each area, pick a go-to person who learns a bit more and handles everyone else's day-to-day questions.

### Step 9. Retire the spreadsheet

This is the step people forget. If the spreadsheet stays open, someone will keep using it "just in case," and three months later you're back to two versions of the truth.

Before go-live, define the criteria for shutting it down, and write them down:

- [ ] One full close cycle is done, and every discrepancy has been explained and fixed.
- [ ] At least two weeks in a row with no new discrepancies in stock, receivables and sales.
- [ ] Every user in the process logs in and works in the system; nobody relies on someone else to pass them data.
- [ ] The reports that used to come from the spreadsheet now come from the system.
- [ ] The process owner and management sign off that the system is the single source of truth.

Once they're met, the spreadsheet goes view-only, moves to an archive folder with the date in its name, and everyone gets notified. From that day on, if it isn't in the system, it didn't happen.

Excel doesn't go away. It's still your analysis tool: export data from the system and slice it, chart it or forecast with it. There's just one rule: never type into Excel something that already lives in the system. If someone does, it's a sign the system is missing something, and that gets requested, not patched with a new spreadsheet.

## A 10- to 12-week plan for one area

Here's what migrating one area, say orders and inventory, looks like at a small or midsize company. The weeks are a guide: if your data is a mess, cleanup runs longer; if the process is simple, the build is shorter and the plan fits in 10 weeks.

| Week | What your team does | What the vendor does | Deliverable |
|---|---|---|---|
| 1 | Spreadsheet inventory and interviews with each area | Sits in on interviews and asks about the exceptions | Spreadsheet list with criticality |
| 2 | Process map and list of business rules | Proposes how the process will work in the system | Current flow, new flow and written rules |
| 3 | Picks the process, names data owners and starts cleanup | Defines scope, screens and reports for the first module | Approved scope and tentative cutover date |
| 4–6 | Cleans up customers, products and prices; reviews progress | Builds the module and demos progress every week | Clean master data and a module ready to test |
| 7 | Tests with real cases, exceptions included | Fixes whatever testing turns up | Punch list |
| 8 | Acceptance testing and role-based training | Loads master data, trains users and hands over short guides | Punch list closed and users ready |
| 9 | Cutover: physical count, receivables cutoff and go-live | Loads opening balances and supports day one | System live with cutover balances |
| 9–12 | Parallel run and weekly reconciliation | Daily support and adjustments | Signed reconciliations |
| 12 | Review of the shutdown criteria | Final adjustments | Spreadsheet archived as view-only |

Notice that your team has work every single week. A migration isn't something the vendor does on its own while the business carries on as usual: the first three weeks and the cleanup depend almost entirely on the people who know the operation. Block out a few hours a week from each data owner, and name one internal project lead.

Notice too that the parallel run takes up a third of the plan. That's not wasted time; it's insurance.

> It's the same scale we work at when Efidis builds [custom software](https://efidis.co/en/servicios/software-a-la-medida): deliveries in 1 to 3 months, one area at a time, with the business still running.

## 7 common mistakes when moving off Excel

1. **Migrating everything at once.** It's the shortest path to a project that drags on and wears out your team before it delivers anything. Start with one process.
2. **Loading dirty data.** If customers are duplicated in the spreadsheet, they'll be duplicated in the system. The difference is that now your team will blame the system.
3. **No owner for each type of data.** If anyone can create customers or change prices, six months later the system looks like the spreadsheet it replaced.
4. **Copying the spreadsheet into the system as is.** Your spreadsheet has columns that exist only because Excel couldn't do anything else. Design from the flow (step 2), not the columns.
5. **Not setting a cutover date.** Without a clear day, half the team records in the system and half in the spreadsheet, and neither one adds up.
6. **Training only the managers.** Managers don't take orders or ship them. The people who do end up learning secondhand, and with mistakes.
7. **Retiring the spreadsheet before the numbers match, or never retiring it.** Shut it off too early and you lose the baseline for catching discrepancies. Never shut it off and you never stop having two versions.

## Frequently asked questions

### How long does it take to move off Excel?

For one area, like orders and inventory, 10 to 12 weeks is a reasonable benchmark, including a month of parallel run. What moves the timeline most isn't building the system but the state of your data and how much time your team can give to cleanup and testing. The whole company moves in stages, one area after another.

### Do we lose our history?

No. The old files are kept in a view-only folder and you can always look things up. What gets migrated is open balances: stock, receivables and pending orders. If you need history to compare sales year over year, it can be loaded in summary form by month, customer and product.

### What if my team doesn't want to change?

Bring them in from the spreadsheet inventory and the process map: the people doing the work know where time gets wasted, and if the system fixes that for them, they'll use it. Train by role with their own data, and leave a go-to person in each area. Then retire the spreadsheet once the criteria are met: as long as it's still open, there will always be an excuse to go back to it.

### Can I keep using Excel?

Yes, for analysis. A good system lets you export whatever you need to Excel: sales, stock, receivables. What you don't want is to start typing into Excel again something that already lives in the system, because that's how the problem starts over.

### What if something goes wrong on day one?

That's what the parallel run is for: during that month the spreadsheet stays current and serves as your backup. Pick a slow cutover date, agree with your vendor on direct support for the first few weeks, and make it clear who to call if something doesn't work. Day-one problems are almost always master data loaded wrong or missing rules, and they get fixed quickly if someone reports them the same day.

## Sources

1. [Panko, R. (2005). What We Know About Spreadsheet Errors (revised version of the Journal of End User Computing article, 10(2), 1998).](https://excel-do.com/wp-content/uploads/2023/09/What_We_Know_About_Spreadsheet_Errors.pdf)
2. [Powell, S., Baker, K. and Lawson, B. (2009). Impact of errors in operational spreadsheets. Decision Support Systems, 47, 126–132.](https://arxiv.org/abs/0801.0715)
3. [JPMorgan Chase & Co. (2013). Report of JPMorgan Chase & Co. Management Task Force Regarding 2012 CIO Losses.](https://elischolar.library.yale.edu/ypfs-documents/454/)
4. [Herndon, T., Ash, M. and Pollin, R. (2013). Does High Public Debt Consistently Stifle Economic Growth? A Critique of Reinhart and Rogoff. PERI Working Paper 322.](https://peri.umass.edu/wp-content/uploads/joomla/images/WP322.pdf)
5. [Ziemann, M., Eren, Y. and El-Osta, A. (2016). Gene name errors are widespread in the scientific literature. Genome Biology, 17, 177.](https://doi.org/10.1186/s13059-016-1044-7)
6. [Abeysooriya, M. et al. (2021). Gene name errors: Lessons not learned. PLoS Computational Biology.](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8357140/)
7. [Smartsheet (2017). Automation in the Workplace: workers waste a quarter of the work week on manual, repetitive tasks.](https://www.smartsheet.com/content-center/product-news/automation/workers-waste-quarter-work-week-manual-repetitive-tasks)
8. [McKinsey Global Institute (2017). A Future That Works: Automation, Employment, and Productivity.](https://www.mckinsey.com/~/media/mckinsey/featured%20insights/digital%20disruption/harnessing%20automation%20for%20a%20future%20that%20works/mgi-a-future-that-works-full-report-updated.pdf)
9. [Bloch, M., Blumberg, S. and Laartz, J. (2012). Delivering large-scale IT projects on time, on budget, and on value. McKinsey and University of Oxford.](https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/delivering-large-scale-it-projects-on-time-on-budget-and-on-value)
10. [Flyvbjerg, B. and Budzier, A. (2011). Why Your IT Project May Be Riskier Than You Think. Harvard Business Review.](https://hbr.org/2011/09/why-your-it-project-may-be-riskier-than-you-think)
11. [Panorama Consulting Group (2026). 2026 ERP Report.](https://4439340.fs1.hubspotusercontent-na1.net/hubfs/4439340/Reports/ERP%20Report/2026-erp-report-panorama-consulting-group.pdf)

## Want to put this to work in your company?

Tell us how your team works today. On a call we look at what to organize first and what it would cost.

- [Book my call](https://wa.me/573214646372?text=Hi%2C%20I%20read%20an%20article%20on%20the%20Efidis%20blog%20and%20want%20to%20see%20how%20to%20apply%20it%20in%20my%20company.)
