IT Glossary
What is a data lake?
A central store that keeps company data in its raw original form — structured or not — so tomorrow's analysis and AI projects still have material to work with.
A warehouse decides in advance what shape the data must take; a lake refuses to decide. That single difference explains most of what follows. A data lake is a large, cheap storage area where you drop everything the business produces — sales exports, sensor readings, contracts as documents, call recordings, website logs, photographs from the production line — in whatever format it arrived, without first designing tables for it. Structure gets applied later, when someone finally has a question worth answering. Firms reach for one because the expensive mistake in analytics is not storing the wrong thing, it is discarding something you did not yet know you needed: the model you want to train in eighteen months can only learn from history that somebody kept. That is the business case in a sentence. The trap is equally well documented. A lake with no catalogue, no ownership and no quality rules becomes a swamp nobody trusts, and the storage turns out to have been the cheapest part of the failure.
Let’s talk about your project
Message us on WhatsApp or send an email — you talk directly to a developer.
office@northdan.com · +40 752 070 247
Why it matters for your business
Nothing gets thrown away
Raw history is preserved even when nobody has a use for it yet, so future questions and models are not blocked by missing years.
Raw material for AI work
Machine learning and document search projects need volume and variety; a lake is where that supply accumulates before it is needed.
One address for every source
ERP, shop, CRM, machines and spreadsheets land in one place, so analysts stop negotiating exports with five different system owners.
Frequently asked questions
Data lake or data warehouse — which does my company need?
A warehouse serves reporting on clean, agreed figures: revenue, margin, stock. A lake serves exploration and machine learning on messy raw material. Most mid-sized firms need the warehouse first because it answers board questions, then add a lake when unstructured sources — documents, images, telemetry — start mattering. Many end up with both, feeding each other.
What does a data lake cost for a mid-sized company?
Storage itself is remarkably cheap: object storage in a European cloud region costs roughly the price of a few dinners per terabyte per month. The real budget lines are the ingestion pipelines, the catalogue and the people who curate it. Expect the engineering around the lake to cost several times the storage bill, every year it stays alive.
How do I stop the lake turning into a data swamp?
Three habits do most of the work: register every dataset in a catalogue with a named owner before it lands, keep the ingestion date and source on every record so lineage is traceable, and set retention rules from day one. A lake without ownership decays quietly, and the people who could untangle it usually leave before anyone notices.
Let’s talk about your project
Message us on WhatsApp or send an email — you talk directly to a developer.
office@northdan.com · +40 752 070 247