Rise up: What the chemical reactions in bread can teach us about data quality

You can’t get a simpler and more satisfying chemical reaction than bread.

A handful of store cupboard ingredients, a bit of manual effort (or a bread maker if you’re lucky) and hey presto – you’ve got a fresh loaf.

But anyone who has taken a disappointingly flat, stodgy sourdough out of the oven will know that it’s not actually that simple at all. Creating the chemical reaction that makes bread rise is a delicate balancing act. To get it to work, you need to pay close attention to what you’re putting in, and the other factors that can impact those ingredients and affect your end result.

The same principles apply to your data strategy. The data you put in is, inevitably, going to influence the results you get out. To get the reaction you want, you need to look at the quality of that data, and how it might be impacted by the processes you put it through, the people who handle it and the technology you use to manage it.

This is where data quality comes in, and it should be a fundamental part of any data project. But to work effectively, data quality needs to go beyond a tick-box exercise and address what a business really needs to get the outcome it wants from its data.

What makes bread rise – and your data strategy work?

For bread to rise, it needs to create carbon dioxide. This reaction occurs when the yeast starts to feed off the sugar and is accelerated when the dough is put in the oven, with the heat expanding and moving the gas to create gas bubbles in the dough.

But there are several factors that can stop that chemical reaction from happening. One, the yeast is too old. Two, too much sugar is added. Three, you miss out an ingredient entirely. Or four, the temperature in the oven isn’t right.

When it comes to data, understanding exactly what you’re adding to the mix is vital to getting the right reaction for your business. A robust data quality exercise should look at those four key areas. Is the data accurate, up to date and reliable? Is there rogue or duplicate data lurking in your databases, that can skew results? Are you missing data that you need? Or is the way that data is being stored and manipulated having an impact on its quality and accuracy?

But it’s not just the data that’s missing that can create a problem. Data that you don’t need finding its way into your system can also create issues. Just like a tablespoon of salt will kill the yeast/sugar reaction (as the yeast releases its water content to the salt by osmosis), trying to process unnecessary, irrelevant data will dilute your data quality and make it harder to get tangible results from your project.