Left or Right? How Data Warehouses Make it Easy to Make Decisions

Deciding between two different courses of action is perhaps the hardest part of running any business. Recently we wrote about how data warehouses can give you insights, trends and information. Everything you need to make informed decisions.

But what sort of decisions do they help with? How do you actually take that insight and make practical choices? In this article, we’re going to talk through nine decisions you might need to make. And how a data warehouse can help.

Why do studios and publishers use data warehouses?

Most hit studios and publishers have a portfolio of games. A data warehouse brings together the data from across them all and converts it into a format they can search through.

With more data, they can spot trends. Knowing the trends, they can fix bugs, optimize their ads, see which ideas perform the best and focus their attention. More attention on the right games leads to more revenue.

Without a data warehouse, it’s like stumbling around a dungeon in the dark, guessing which way leads to the treasure. A data warehouse gives them true sight: they can see things for what they truly are.

And with our Player Warehouse – a data warehouse specifically for game developers – you’ll see the same GameAnalytics interface that you’re familiar with. You can find out a lot from our regular tools, but the Player Warehouse is the next step to leveling up your analysis.

Nine insights you can pull from a data warehouse

Once you’ve set up your data warehouse, it’s time to start asking some practical questions. But where do you start? With so much information at your fingertips, it can be a bit daunting. So let’s look at nine ways you can use your data warehouse.

Decision one: Which paid items can I promote? And which can I scrap?

With a data warehouse, you can see which In-App Purchases bring you in the most revenue. And which ones aren’t. Once you know that, you can decide which ones to show players and which ones to ditch.

We recommend running an A/B test to see whether it’s the purchase or another factor. It could just be that you have too many options, which can frighten people off. (There’s a good TedTalk on this: Sheena Lyengar explains how too much choice can harm sales.) It’s also possible that it’s just in the wrong place. Placing a poor offer next to a good one, makes the good offer look even better.

It’s worth noting that in our web tool, we give you an analysis of any A/B tests you run. But with our Player Warehouse, you can see every player from an experiment and come up with your own conclusions.

Once you’ve run your test, you can add your most popular purchases to your onboarding and you should see a bump in your Average Revenue Per Paying User and your Lifetime Value.

What queries should I run?

SELECT item_id, is_converting as new_buyers, COUNT(*) AS purchases

FROM project_id.dataset_id.payment_checkpoint

WHERE checkpoint = “2020-01-01”

AND game_id = 123456

GROUP BY item_id, is_converting

ORDER BY purchases DESC

Decision two: What should I do with my whales?

Your whales are extremely important to your bottom line. With a data warehouse, you can see how many you have, how often they spend, and what precisely they’re doing in your game.

Once you know these spending habits, you can decide exactly what to show them. You can find out which items or packages they prefer. For example, do they like to buy the largest bundle or lots of small bundles? How often do they take up an offer? Is it better to have them make one big splash or entice them to spend a little every month?

What queries should I run?

SELECT player_id, SUM(amount_usd) AS total_revenue, AVG(amount_usd) AS avg_purchase_revenue

FROM project_id.dataset_id.payment_checkpoint

WHERE checkpoint >= “2021-01-01” AND checkpoint < “2021-01-08”

GROUP BY player_id

ORDER BY total_revenue

Decision three: Should I cross-promote or use ads?

While a popular game with high retention might draw in a lot of Lifetime Value from your players, not every game is going to draw in a huge amount from ads.

These smaller games – ones that bring in little revenue – might still bring in a lot of new players. You can use those games to move players over to a more in-depth game which you know has a better average revenue per user.

In general, it’s best to cross-promote on games where you aren’t seeing much revenue or on games where too many adverts would disrupt the flow and lose you the player. Instead, you push these players through a funnel to a game you know has high returns.

You might also decide to cross-promote certain groups of players to a game more suited to their spending style. Do you have a group of players who tend to use in-app purchases, but they’re in a game where you don’t have many? Guide them to a different game of yours.

Your ads are your income. But cross-promotion is like a sieve, separating people to the places where they’ll be most valuable.

What should I do?

The best way to figure out whether you should cross-promote or use ads is to either plug your data warehouse into a machine learning algorithm or run a bunch of A/B tests. For example, you could run an A/B test to see where to place your ads.

Decision four: Which players should I show more IAP offers to?

Not every player reacts in the same way to offers and In-App Purchases. With the right data, you can start creating groups and divide your players into different categories (this is especially important for your whales). You can change which offers you show them. Perhaps you give offers on cosmetic items to one group, while you show another loot boxes in bigger bundles.

What queries should I run?

SELECT game_id, user_id AS user_with_potential_to_spend_in_that_game

FROM `project_id.dataset_id.player_checkpoint`

WHERE checkpoint = “2021-01-01”

AND revenue_usd IS NULL

AND user_id IN (

SELECT DISTINCT(user_id)

FROM `project_id.dataset_id.player_checkpoint`

WHERE checkpoint = “2021-01-01”

AND is_paying_today

)

This query shows you players that were active today and made an in-app purchase in at least one of your games. You can then see which other games they play and did not spend in, so that you can make them an offer.

Decision five: Which players should I show more ads to?

Similarly, if you create groups of players you can spot which players drop off your game when they see ads. When they next download a game from you, you can make sure to show them fewer adverts and keep them around for longer. You could even show them the option to remove the ads.

Likewise, you can add a player to a group that doesn’t make In-App Purchases. With this group, you can deliberately show them more adverts, show them more offers encouraging them to spend, or add rewards for their patience.

Decision six: What should I do with my weakest games?

Don’t write off a low-earning game, just because it’s not creating revenue. Perhaps it has other uses. For example, it could help you discover whales or introduce your players to your other titles. You can use these games to find your best players and even drill down further, creating a group: “Players who play more than one of our games.”

You can also look at why those games are performing poorly. Are there too many ads? Do they have different features? Is it the genre? Learning this, you can experiment and make tweaks to see what works. You can then use this knowledge on your better-earning games to really see a difference.

Decision seven: Which players are likely to give good reviews?

Even low-spending players can have value in other ways. When creating your groups, look for players that often give good reviews. (Often these are the ones with the highest playtime, installed a long time ago, spent a lot of money or have had no bugs flagged.) You can now create a new group and prompt them to leave you reviews.

If you need to give them more incentive, give them a freebie for their time. They’ll see that you value them and are even more likely to leave a positive review.

Decision eight: Which levels are best for my tutorial?

Spotting a tricky level can help you see whether you need to improve your tutorial. It could be that you didn’t explain a concept well enough or that your players aren’t remembering a certain mechanic. It could even be a bug that needs fixing.

If you see a problem during your onboarding, it’s best to remove it and stick it in for review. Rather than leaving it while you try to solve the problem. A quick decision now can help increase the playtime and retention.

What queries should I run?

SELECT progression_1, progression_status, COUNT(DISTINCT(player_id)) AS players

FROM project_id.dataset_id.progression_checkpoint

WHERE checkpoint >= “2020-01-01” AND checkpoint < “2020-01-08”

AND game_id = 123456

GROUP BY progression_1, progression_status

ORDER BY progression_1 ASC, progression_status DESC

Decision nine: Which countries should I spend my advertising on?

You’ll find that some countries will give you players who spend money, while others are useful to run tests. With the right searches, you’ll find out which gives you the best Lifetime Value. These are the countries you want to take the least risks with and show the most ads to.

Meanwhile, countries with low Lifetime Value are useful for soft launches, releasing major patches to make sure they’re working and big experiments.

What queries should I run?

SELECT checkpoint, country_code, game_id, COUNT(DISTINCT user_id) as DAU

FROM project_id.dataset_id.player_checkpoint

WHERE checkpoint >= ‘2020-01-01’ AND checkpoint < ‘2020-01-08’

AND checkpoint = install_ts

AND country_code IN (‘US’, ‘GB’, ‘SA’, ‘CN’, ‘FR’, ‘AU’, ‘BR’)

GROUP BY checkpoint, country_code, game_id

ORDER BY checkpoint DESC

Get started with our Player Warehouse

We designed our Player Warehouse specifically with developers in mind. It brings together the data from all your games, connects with our other tools and lets you link up with third-party applications. Get started and make the most of your games.