Dirty data is the mother of all screw-ups

March 23, 2023
Chelsea Africa
min read

Despite the mobile gaming industry taking a major knock in 2022, there were still almost 90 billion mobile game downloads for the year. This is according to data.ai’s State of Mobile 2023 report. To give you an idea of how mind-blowing this number is - that’s an average of 11 game downloads per living person on Earth. Okay, so it’s impossible for every human to have downloaded a game last year but our point is - that’s a lot of data, especially the more you segment it!

Data, as we know, is super important for all businesses since it helps pin-point opportunities and predict trends. So, in an industry with that much data floating about, you can only imagine how important it is to maintain the quality of data used to inform strategic business decisions.

To understand why clean data is simply non-negotiable, we need to know what constitutes dirty data. These are the two things that = sh*tty data:

  • Mistakes - spelling errors, duplicates, missing fields, incorrectly name fields etc.
  • Outdated inputs - data that is no longer valid and has not been updated accordingly

Dirty leads to filthy forecasting, and that means:

  1. Wasted resources

    Fixing mistakes, removing duplicates, filling in missing data etc. means money and time wasted. These are resources which can be better spent making critical decisions about game titles, implementing impactful features or shifting to profitable channels in order to scale.

    Here’s a practical example of wasting resources:

    Mobile Gaming studios have pretty expensive internal resources (we’re talking Financial Planners & Analysts, Finance Business Partners, Data Scientists etc.) who are there to share insights with management teams in order to drive positive ROI. As a studio, you’re paying these incredible people (that’s not sarcasm - we work with some of the best) to crunch the numbers and spit out forecasts, right?

    But instead said pretty expensive resources are spent fixing mistakes (we’re human, so this is not farfetched even at the highest level), or building a process to avoid it. So, when do they do the most important part of their jobs: analyse and share insight? They either don’t - or they do a rushed analysis because they have too little time left.
  1. Bad decisions

    Since data is used to derive insights from, incorrect data will lead to ill-informed decisions which could have catastrophic implications.

    Here’s a practical example of a bad decision:

    A studio with bad quality data may decide to invest resources in developing a new game feature or monetisation model based on their available cohort data. However, after going live with the feature or monetisation model, they realise that there is zero to no uptake from their cohorts - even worse, the new feature has led to more churn. The cause? The data that informed their decision was sh*t - and now it’s too late, the time and money used to develop the feature or monetisation model is wasted.
  1. Missed opportunities

    It’s important to note that bad decisions are not the same as missed opportunities. Where bad decisions are based on incorrect information, missed opportunities are more often than not caused by not having access to such information at all.

    Here’s a practical example of a missed opportunity:

    We’re bringing this example closer to home and flipping the narrative. Imagine if our Co-Founder and CSO Angus Lovitt didn’t manage to get within 5% of his predicted revenue of Candy Crush, after already convincing the board at King to spend $350 Million on marketing based on his promise to achieve it.

    Because of clean data, which led to an accurate forecast, Angus empowered the board to make a data-driven decision that allowed them to capitalise on an opportunity as opposed to missing one.

Here are our recommendations for maintaining clean data:

  1. Ensure the entry point is as clean as possible

    Prevention is better than cure. Have a look at the entry point of your data and ensure it’s coming in as clean as possible to begin with. This may take some upfront work, but it will save you and your team in the long run.
  2. Standardise and document processes

    Agree on cleaning processes - and make sure this is shared with everyone who will handle the data. Not only does this remove the risk of bad quality data, but it also allows for knowledge sharing which is essential for business continuity and individual growth.
  3. Regularly audit processes

    Regular auditing helps to identify and fix errors quickly, preventing them from accumulating and leading to more significant issues down the line. It also ensures that standardisation, mentioned in the point above, is being followed.

    Put together an audit schedule - perhaps you agree to look at things every six months or quarterly. If you’re gathering large amounts of data over short periods of time (say daily, or weekly), you may need to schedule more regular check-ins.
  1. Choose the right tool

    There are many data management and analysis tools available on the market, and choosing the right one can help to improve the quality of data and streamline data management processes. The right tool can also help to automate data processing and management, reduce errors, and provide better insights into the data and reduce employee burnout.

    Studios should consider factors such as the size and complexity of their data sets, the need for real-time data processing and analysis, and the level of automation required.

With mobile game downloads predicted to continue on an upward trajectory in 2023, and subsequently the amount of data too, studios should triple-check their data integrity. Even more so, due to the storms that lie ahead for the mobile gaming industry, studios cannot afford (literally!) the cost of making bad decisions as a result of bad data right now.

Related Articles