The data normalization process takes place at the very beginning of your experience with Libring, mainly during the trial period. However, normalization is an ongoing process, as you want to make sure that you are constantly retrieving meaningful data to your business. Although it may seem tricky sometimes, it’s good to remember that translating data into an organized list, getting rid of invalid records and standardizing naming conventions will make your life much easier.

We have come up with a few tips and tricks to keep you on top of the data normalization process and stay ahead of your incoming data.

1. Pick Simple Aliases

Every company is different. Maybe it’s easier for you to put down the full app name in order to organize everything, or maybe you have a nickname you use for it. By using names and terms you can easily identify, the data normalization process becomes a lot easier to understand.

You need to be consistent. The last thing you need is the same app popping up in multiple line items.

2. Watch Spelling!

When is comes to standardizing naming conventions, you need to take things like capitalization, spacing, spelling, and numbers into consideration. Nothing is more annoying than having a single apps data pop up in separate line items because half the data is coming in as “snake” and the other half as “Snake”.

What if your apps have sequels to them? Again, naming is key. Here is an example: you have data coming in from three sources: Snake, Snake 2, and Snake v2. However, is Snake 2 and Snake v2 the same game? Once you figure out that you need to make the necessary changes so that all your data will input under one line item.

3. Use the Pivot Table to spot inconsistencies

One of our favorite recommendations is to use our Pivot Table feature to spot irregularities. By dragging and dropping “App” and “Connection” into the table, you can easily identify which apps are not normalized and the connection generating it. This makes it easier to set the rules you want and change information within networks so your data comes in normalized.

Data Normalization

4. Get rid of the descriptors at the end of your app name

We have plenty of dimension columns, and even 4 custom columns if you need them. So you can make the distinction between the platform iOS and Android in the column Platform, and get rid of the descriptor that is tacked onto the end of your app name. By doing this, the system will summarize your metrics accordingly and you can analyze the performance by app and platform separately making for easier analysis down the road.

This also goes for dimensions like ad type. Trust us, just get rid of the descriptors.

5. Set rules to help normalize incoming data

When you are making adjustments in normalization, they will only be reflected in your current set of data. However, there is a way to create rules to have your changes reflected in your future line items as well. Rules allow you to save the logic behind your naming convention structure. After you’ve created a rule, all new items will be automatically adjusted according to that rule. Keep in mind that the order of the rules can impact how the aliases are applied.

For Example: You have two original items, one with “snk ios” and another with “snk 2 ios”, you want to set their app names respectively as Snake and Snake 2. A good strategy would be to start by first creating a rule for the more generic case and then to create a rule for the more specific one.

• Rule 1: When App=”snk ios” change to Snake
• Rule 2: When App=“snk 2 ios” change to Snake 2

Be careful when applying conventions to rules as well. If you just type App=”snk” you could change all of the line items that include “snk”, even though they are not the same app. In these cases, the order of the rules matter. A rule to differentiate “snk 2” should come before a rule for just “snk”.

6. Set up Custom dimensions to have more granular groups

Use the 4 custom columns to add extra information to your data to extract deeper insights. By adding more info at the normalization stage, you are opening the door for the use of this info in the filtering option of other features and on the Pivot Table for easier reports.

7. No need to normalize Geo | Country info

Libring deals with Geo/Country data normalization internally so you don’t need to worry about it. It doesn’t matter the way this information is formatted. Our smart data collection structure will organize everything automatically for you.

8. Remember to apply our Smart Filters

Our Smart Filters enable you to easily spot any empty cells, cells with original names, cells that have been normalized with a rule, and cells that have been manually changed. You can refer to the Dimension Tokens within Libring Edge to go even more granular in your filtering. 

Smart Filters can be applied if you use the following syntax:

app= Restricts search to App Name field
platform= Restricts search to Platform field
ad_type= Restricts search to Ad Type field
ad_format= Restricts search to Ad Format field
custom_1= Restricts search to Custom 1 field
custom_2= Restricts search to Custom 2 field
connection= See you will see only rows for a specific connection
connection:id= Filter the rows by the matching connection ID
connection:original= Filter the rows by the matching Connection Name
connection:alias= Filter rows by matching Connection Given Name
connection:segment= Filter the rows by matching segment (ad network, mediation, attribution, etc.)
id= Filter the row by the line item’s matching unique ID
enabled= all row enabled
disabled= all rows disabled

To see them all listed out on the platform. Simply click on the questions mark (?) on the right side of the Normalization search bar.

Don’t hesitate to contact us if you have any doubts about data normalization or any other feature! We’ve also created this video to guide you through the process within the platform.


*Originally posted in March 29, 2017 and updated in March 7, 2019