Avoiding duplicate data entry: typical causes

Four causes

• Two systems, no connection. The classic: the shop knows the order, the merchandise management system does not.
• The system cannot do something, so a spreadsheet appears alongside — and is maintained too.
• Information arrives as free text: by email, by telephone, on paper. Someone transcribes it.
• An evaluation needs data in a form that no system provides. So someone assembles it anew every week.

None of these causes is an organisational failing. All of them arise because someone bridges a gap — and that is precisely why they do not disappear when you forbid them.

What it costs

The typing time is the smaller part. The consequences are more expensive:

• Two versions: when they differ, someone has to decide which one applies — and only someone who knows both can do that.
• Transcription errors: a wrongly transferred digit often goes unnoticed until it takes effect.
• Delay: between entry and transfer there is a waiting time in which the second system is misinformed.
• Dependence on one person: whoever does the transfer is the only one who knows its quirks.

Four ways out

• Structure the input: what arrives as free text today arrives through a form. Often the cheapest solution — it removes the cause rather than the consequence.
• File exchange: one system writes a file, the other reads it, usually overnight. Robust, possible anywhere, up to a day old.
• Interface: the systems talk to each other directly, with validation and feedback. The cleanest way, requires coordination.
• Abolish the second entry: sometimes the second system does not need the detail at all — it has just "always had" it.

The fourth way is rarely examined and is surprisingly often the right one. The question is not "how do we transfer this", but "who needs this detail, and for what".