Continue to WhatsApp?

The chat opens in WhatsApp, in the app or in a new tab. You leave our website, and data such as your IP address is sent to WhatsApp (Meta).

DATA AND IMPORTS

Transfer data between systems: 4 ways to stop typing it twice

The orders are already in the web shop, yet someone types them into the ERP line by line: customer, item, quantity, price. In the back of their mind is the nagging worry that a digit got swapped somewhere and only a customer complaint will reveal it. In many offices, transferring data between systems still means retyping it.

Office worker typing data into a laptop while a colleague holds a clipboard with paperwork
Photo: Vitaly Gariev on Unsplash

Moving data between systems by hand costs you twice. Once in working time, every single day. And again in errors that later trigger follow up questions, credit notes or correcting entries. Often the whole process also depends on one colleague, and it stalls when she is on holiday.

This article covers the four ways to get data into the next program without retyping it, the order in which to check them, and how to set up a CSV handover that runs just as cleanly every week. You also get a worked example, the common mistakes and clear criteria for when automation is not worth it.

Quick answer: how to transfer data between systems

Transferring data between systems means a record that already exists in one program reaches the next one without being typed again. There are four routes. Check them in this order, because the first one that fits is usually also the most stable.

  1. Check what your software already does

    Many programs already include an import, an export or a ready made connection to common software. A look at the manual and settings often makes any further project unnecessary.

  2. Use an API

    An API is a documented access point through which programs exchange data directly. It is the cleanest route, but it needs the right licence, permissions and someone to connect it.

  3. Hand over a file

    System A exports a file, usually CSV or Excel, and system B imports it. This works with almost any software, as long as columns, formats and checks are clearly defined.

  4. Automate the user interface

    If there is neither an API nor an import, software can mimic typing on screen. This is called RPA (robotic process automation). On Windows 10 and 11 you can try it yourself with Power Automate for desktop, but scheduled and unattended runs need a paid licence. RPA is the last resort, because a single changed menu can stop it.

The four routes compared
RouteRequirementStabilityTypical use
Built in import or exportFeature exists in the programHighMaster data, bookings, price lists
APIDocumentation, licence, credentialsHighOngoing transfer, many records
File (CSV, Excel)Export in A, import in BMedium to high, depending on checksDaily or weekly handovers
User interface (RPA)Stable screens, fixed stepsLow to mediumLegacy systems without import

Take stock: where does double data entry happen?

Double data entry means a person copies information that already exists digitally from one program into another. As long as nobody names these handovers, they look like normal work. So write them down before you talk about technology. Half an hour with the person who types it today is usually enough.

  • Source and target: Which program does the data come from, and where does it go? For example web shop to ERP, email to Excel, or Excel to accounting software such as DATEV (Germany), BMD or RZL (Austria).
  • Leading system: Which program is right when in doubt? Without this rule, changes overwrite each other.
  • Key: Which field identifies the same record in both programs? Typically customer number, item number or order number.
  • Volume and frequency: How many records, how often? Ten orders a week need a different solution than two hundred a day.
  • Special cases: What happens today when a field is missing or a customer is new? These cases decide the effort later.

The result is one page that states clearly what is done by hand today. With it you can ask any vendor or service provider precise questions instead of commissioning a vague project.

Five questions for your software vendor

  1. Is there an import for exactly this data, such as orders, customers or bookings, and where is the format described?
  2. Is there a documented API, and is it included in our version and licence?
  3. Which fields are mandatory, and how are existing records recognised?
  4. Do we get a log of rejected records with the reason?
  5. Is there a ready made connection to the other program that we only need to switch on?
Office worker typing data into a laptop while a colleague holds a clipboard with paperwork
Photo: Vitaly Gariev on Unsplash

How do you hand over data cleanly with a CSV file? Step by step

If the target program has an import but no direct link to the source, a file is the pragmatic route. Here is how to set it up with Excel and Power Query. Power Query is the tool under Data > Get Data, included in Excel 2016 or later for Windows and in Microsoft 365, including on the Mac.

1. Define the field mapping

Write a small table showing which source column goes into which target field. Note mandatory fields, formats and rules. The target program's import description lists the fields it expects. Accounting software is strict here: DATEV imports bookings and master data in the DATEV format, a CSV file with semicolons and a fixed header line. A plain Excel list will not do. RZL's data import manual describes its own interfaces for bookings and master data in the Windows-1252 character set, and RZL also accepts bookings in DATEV format. For BMD, ask BMD or your tax adviser for the import description for your module. The vendor always defines the format, the fields and the module you need.

2. Load the source file with Power Query

Do not open the export by double clicking it. Use Data > From Text/CSV, choose the delimiter and the file origin, which is the character encoding, then click Transform Data. If umlauts look wrong in the preview, the file origin is wrong, usually UTF-8 instead of Western European (Windows) or the other way round. Set columns such as item number or postcode to the data type Text. This keeps leading zeros. If a single file is already broken, say with “Müller” instead of “Müller” or everything in one column, our free Repair CSV tool shows what is wrong and converts the character set, delimiter, decimal sign and date format to suit the target. The file stays in your browser.

3. Reshape values for the target

Rename columns to what the target program expects. Read dates and amounts using the source locale, the regional setting for numbers and dates. Then convert them into exactly the format the target requires. Power Query records every step. Save the next export under the same name in the same folder and click Refresh All. Save the result via File > Save As as CSV, in the encoding the target program expects.

4. Test with a few records

Import five to ten rows first, including one deliberate special case. Check in the target program that everything landed in the right field. Only then import the full file. An error in ten rows is fixed in minutes, one in a thousand rows costs an afternoon.

5. Reconcile after the import

After importing, export a list from the target program and compare it with the source by key. Our free Excel comparison tool shows missing and changed rows at once, and the files never leave your browser. How to merge two lists by a key column is explained in matching Excel lists.

Example: web shop orders into the ERP

Part of the field mapping (sample data)
Field in shop exportField in ERPRule
order_idExternal document numberMandatory and unique. Numbers already imported are skipped.
customer_emailCustomer numberLook up by email address. No match: flag as new customer for review.
skuItem numberImport as text, keep leading zeros. Unknown number: row goes to review.
qtyQuantityWhole number greater than 0.
price_gross, tax_rateUnit price netDivide the gross price by 1 plus the VAT rate, for example 1.20 at 20% (AT) or 1.19 at 19% (DE). No VAT rate: review.
created_atOrder dateConvert 2026-11-12 to 12.11.2026.

The first row matters most. Without the rule on the external document number, an export imported twice by mistake would duplicate every order. Repeatable runs without duplicates are the core of any handover. For the employee this means she no longer checks every line, only the few the system flags.

Common data import mistakes and how to avoid them

  • Leading zeros disappear: Load item numbers and postcodes as text, not as numbers.
  • Decimal separator and date order mixed up: An export in US format delivers 1,234.50 and 11/12/2026, while German and Austrian software expects 1.234,50 and 12.11.2026. Set the source locale in Power Query on purpose.
  • Umlauts turn into garbage: “Müller” becomes “Müller” when the encoding does not match. Ask whether the target program expects UTF-8 or Windows-1252 (ANSI), and save accordingly. A file that is already garbled can be fixed with Repair CSV.
  • Duplicates on the second run: Every handover needs a unique key and a rule for what happens on a match.
  • No leading system: If the same address is edited in both programs, chance decides. Define where changes are made.
  • Rejected rows go unnoticed: Many imports skip faulty rows silently. Read the import log and count records before and after.
  • It depends on one person: If only one colleague knows the steps, the process stops during holidays. Write the steps down on one page.

What does double data entry cost, and when is automation not worth it?

Two minutes per transaction sounds like nothing. A simple calculation with cautious assumptions: 15 transactions a day, 2 minutes each to transfer. That is 30 minutes a day. Over roughly 220 working days a year, it adds up to 110 hours. At an assumed internal cost of €38 per hour including overheads, that is about €4,200 a year. Errors, follow up questions and corrections are not included.

Run the numbers for your own case in the office work cost calculator in two minutes. It works per month: for this example, enter 275 transactions at 2 minutes each. Treat the result as a guide, not a promise, because an automated process also needs some review time. It does give you a solid figure for the conversation with management or a service provider.

High bay warehouse with boxes on pallets whose stock levels must match in the ERP and the web shop
Photo: Arno Senoner on Unsplash

Automation is often not worth it when:

  • there are only a few records a month and copying them causes no errors,
  • you plan to replace one of the programs in the next few months anyway,
  • the rules change every time and someone has to decide case by case,
  • the target program already solves the task with an existing feature.

When an office helper makes sense

A file handover with Power Query goes a long way, as long as one person triggers and checks it regularly. It gets harder when several sources come together, when data has to be read from emails or PDFs, or when special cases need to be sorted out reliably.

That is where an office helper takes over exactly this one handover: it collects the data, maps the fields, checks the agreed rules and passes the result to your program through an API or import. Anything that does not fit lands on a list for your team, and approval stays with you. Your software stays, and the process keeps running when the colleague in charge is on holiday. See how this works on transferring data between programs automatically. If your data comes from invoices in a mailbox, read invoice processing without double entry.

For an assessment we need just three things: which program the data comes from, where it needs to go and how many records a month. Put them into the workflow check and you get a personal, free assessment within 1–2 working days. If your software can already handle the handover, we show you where. If an office helper makes sense, we quote a fixed price from €1,990 excl. VAT, and small projects are usually done in 3–7 days once the use case is agreed.

Sources and further reading

FAQ

Quick answers.

What is double data entry?

Double data entry means typing information into a second program even though it already exists digitally in a first one. A typical case is an order from an email or web shop that someone retypes into the ERP. It costs time on every record, is a common source of errors and can often be replaced by an import.

How do you transfer data between two systems without an API?

Usually with a file: the first program exports a CSV or Excel file and the second reads it through its import. In between you define the field mapping and checks, for example with Power Query. If there is no import either, only user interface automation is left, which breaks more easily when screens change.

What is the best way to sync data between two systems?

First decide which system is leading and which field identifies a record in both. Then use the most stable route available: a built in connection, an API, or a file import with checks. Two way sync is only needed if the same record is edited in both systems, and it needs clear rules for conflicts.

How can Excel transfer data automatically from one file to another?

The most robust way is Power Query under Data > Get Data > From File > From Excel Workbook. The query reads the source file, reshapes the columns and loads the result into the target workbook. As long as the source file keeps its name and location, one click on Refresh All picks up new data.

What are typical manual data entry errors?

Swapped digits, missing leading zeros, wrong decimal separators, wrong date order and records entered twice. Many of these errors only show up later, for example in an invoice, a payment reminder or the stock count. A fixed key and a reconciliation after each transfer catch most of them before they cost money.

LET'S TALK ABOUT YOUR WORKFLOW

Write to us. We reply personally.

A question, an idea or a task that costs time every week? A few sentences are enough. You get an honest answer from the people who also build your office helper.

Ivan, Lukas & JanReply within 1–2 working days

  • Personal reply
  • No obligation
  • Based in Vienna, GDPR aware
Photo of Ivan
Rather talk directly?Ivan, your contact person+43 676 59 40 027office@pebblebyte.comMon to Fri 08:00–18:00
Have a specific task? Get your free workflow check

Your message

Please do not enter passwords or confidential customer data. We only use your details to handle your request. No newsletter, no sharing for advertising. Details in our privacy notice.