In the world of data, from BI to data science to AI, everyone has heard the phrase ‘garbage in, garbage out’ at some point. In fact, we hear it far too often. It usually comes across as a shrugging observation: the data wasn’t good, but we worked with it anyway, and that’s the end of the matter.
Yet that conclusion isn’t right.
So that’s that
The way ‘garbage in, garbage out’ is usually used is as an end point. The data was messy, so the result is messy too, and as soon as the report, dashboard or forecast has been delivered, responsibility has been handed over. The problem is acknowledged, and then let go.
That attitude seems harmless. It gives a team the space to carry on working without having to reopen the discussion. It isn’t, because the output it justifies doesn’t disappear. It re-enters the system somewhere else.
Output is always someone else’s input
No output ever remains where it is created. A report is read by a manager. A dashboard drives a decision. An AI forecast triggers a follow-up action. The ‘garbage out’ of one process is, almost always, the ‘garbage in’ of the next.
That is why the reversal is at least as important as the original. Garbage out is garbage in.
From operational to strategic, and back again
Take an operational report with incomplete figures. That report feeds into a tactical overview, compiled by someone who can no longer see the underlying mess. That overview, in turn, feeds into a strategic decision, taken by someone even further removed from the source.
At every step, the data becomes a layer more abstract, a layer more convincing, and a layer less visibly incorrect. No one in that chain can recognise the original rubbish. Everyone sees only the input they themselves receive, and takes it seriously.
Why nobody feels a sense of ownership
That is precisely the problem with the way we use the term. Whoever produces the output does not feel responsible for what happens to it once it has been delivered. They have done their bit.
Whoever receives the output, in turn, assumes that it has already been checked by the person who supplied it. They have no reason to doubt it, especially when that output is presented in a polished manner. A sleek dashboard or a well-organised report looks reliable, and that appearance is quickly confused with the reliability of the data behind it. It is the form that convinces, not the content.
Caught between these two assumptions, quality falls by the wayside. Nobody is lying, nobody is being negligent, and yet nobody checks.
AI speeds up the chain, it doesn’t shorten it
With AI, that effect becomes greater, not smaller. A model uses output from one system as input for the next, often without a human intervening to assess that intermediate step. Agentic AI even takes that next step autonomously, based on what it has just produced itself.
So the ‘garbage’ not only travels further through the organisation, it also travels faster, and with less chance of anyone noticing it along the way before it serves as input somewhere else.
A warning, not a licence
‘Garbage in, garbage out’ should therefore not be an excuse to use poor-quality data anyway, on the assumption that once the output is produced, you’re done with it. It should be a warning that travels with that output, throughout the entire process.
Anyone who realises this will ask an extra question with every delivery. Not just whether the data is good enough for this purpose, but also where this output will end up next, and with whom.
Garbage out is garbage in!
The next time someone shrugs and says ‘garbage in, garbage out’ to justify poor-quality data, the answer is simple. The conversation doesn’t end with the output. It starts again there, for someone else, at the next level.
Garbage out is garbage in. Until proven otherwise.
Want to know more about data quality? Follow Ruud on LinkedIn for updates on data. Also read our other blogs on MAD-Quality.
