This year at the Videoweek Villa in Cannes, Pickaxe co-founder Andrew Grosso discussed becoming a personal trainer for data. What we’re really talking about when we say “personal training” is building fundamentals, setting benchmarks, and staying consistent. A personal trainer does more than get you in shape – they help you build lifelong habits and practices to stay healthy in the long term.
So is your data in shape? Here are some of the indicators that can help you find out.
Does your data show up on time?
Do you know when your data is coming in – and just as importantly, when it’s not?
Having a predictable schedule around your data is key to ensuring that it’s available when you need it, where you need it. For example, a Netflix analyst evaluating the most watched programs or the completion rate for a particular series doesn’t need real-time data; it’s probably fine for them to review a data set that was updated yesterday or at the conclusion of the week. But a Netflix user finishing a show and seeing the “Watch Next” recommendation section in the app doesn’t want to see a recommendation of something they’ve just finished watching, either. In this case, the data needs to be updated in near real-time. Both use cases are using viewing history, but they each need it in very different granularities.
Figuring out how often you need your data, and setting a process in place to make sure it’s arriving when you need it, is a great step to take to make monitoring your data pipeline easier. If something doesn’t arrive on time, or it arrives in a different form than expected, it will enable you to catch and resolve problems right away.
Is your data in the shape you expect it to be?
Sometimes, your data may arrive in a format that doesn’t match the configuration of your existing database. Or you’ll get unexpected values in the existing configuration. In nerd terms, you might get a new column or suddenly see unexpected values in an existing one.
Maybe your company has introduced a new promo code, and wants to track the behavior of customers who have used it. Or maybe something bigger happens – perhaps your company has expanded into a new market, and wants to measure new information related to that. That new data coming in will have fields that don’t match the existing data you have.
You might also receive an unexpected volume of data. If a recent event has happened that has brought a lot of attention to your company (like, for instance, a Super Bowl ad or a viral TikTok), you may end up with a massive influx of data events to track. If you normally get 1000 rows of data every day, and suddenly you’re hit with 1 million, that’s a situation that will need to be addressed.
Part of keeping your data in shape means making sure that a system is in place to notify you about new fields, values, and anomalies. Being conscious of what shape your data is supposed to be in, and sudden changes, can help you figure out how to address shifts that might cause snags when it’s time to use that information.
Are your tools and use cases up to date?
Speaking of the shape your data is in – are your use cases up to date?
As time goes on, goals change, and so do metrics. For example, when a dashboard is initially put into place, it may be designed around monitoring a data category, but it might prove to not be as informative as a different category. If strategy changes, numbers that were important at one point might become irrelevant and that dashboard might become outdated.
Periodically auditing data dashboards and other tools to make sure they’re still relevant is key to making sure they’re useful.
Staying in shape (so you’re ready for anything)
If you’ve got a baseline level of fitness for your data, you’re better able to handle challenges when they arise. Whether you’re dealing with the big changes that come with a product launch, recovering from a disaster in your system, or handling a sudden influx of data from a large event, establishing maintenance practices will make it easier to handle pressure. By getting your data in shape today, you’ll be able to achieve even more with it tomorrow.
Ready to get your data into a fitness regimen?
If you’ve got a baseline level of fitness for your data, you’re better able to handle challenges when they arise.
