Intro
Measuring data health is a process of determining whether your setup is correct, and therefore whether your data is accurate with nothing missing.
With good data health, you can safely rely on the insights from your data discovery. With poor data health, you may need to apply some fixes to improve it.
Market Health Indicators
There are three health status indicators which rate a market's data health:
Status | Integration setup | Match rate | Tracking rate |
Green ✔️ (Optimal) | Successfully received data from Advertising and Analytics Assets in the last 7 days | Match rate from Facebook, Google Shopping, and Analytics Assets is ≥90% | Tracking rate is ≥90% |
Orange ⚠️ (Moderate) | Missing either Advertising or Analytics data | Match rate from Facebook, Google Shopping, and Analytics Assets is <90% and >65% | Tracking rate is <90% and >50% |
Red ❗ (Needs Attention) | Has no inventory | Match rate from Facebook, Google Shopping, and Analytics Assets is <65% | Tracking rate is <50% |
How to Check Data Health
Browse to the Markets and Assets menu to see the in-depth data health status of each market through a tooltip when hovering over the respective market.
In the case of unhealthy data, following automatic suggestions to improve data health can be useful to establish the reliability of a market's data, and therefore increase the validity of corresponding data insights.
Additionally, there are health status indicators anywhere that markets are displayed, such as in Segments Dashboard or Product Insights.
Data Health Metrics
There are two main metrics to monitor:
Match rate
Tracking rate
Of the two, match rate is far more important. When the match rate is wrong, it can influence the entire data set and therefore the calculations. A low tracking rate merely means that it is not possible to separate dynamic and static revenue and connect revenue with the campaign, but that is the only limitation and it is still possible to work with the data.
Match Rate
Match rate shows how much data was able to be paired with products in your inventory. It is calculated as a percentage from all data downloaded (spend, revenue, or other metrics) that was able to be paired with products. This value can be shown on the asset level, market level, and company level. The difference between a hypothetically perfect 100% rate and the actual match rate is a matter of data loss and can happen for several reasons:
Level has an incorrect Product Hierarchy Level Set: This means that the pairing was set to an inventory’s columns where the IDs either don’t match at all or don’t match in the best possible way. It can be caused by either an incorrect pairing setup or an incorrect inventory setup.
Product is not in the Inventory: This means that a statistic is trying to be paired with the spend or revenue for a product which is not in the inventory. If the product is not in the inventory then it was never presented in the inventory source feed. Usually, that is due to not all of your products being included in the feed that was selected as the source of the inventory. However, other reasons can cause this issue, such as an inventory feed not working properly or inconsistent data on your side.
Data loss means that you are missing some data about product performance. It can cause poor calculation of segments and Product Insights, depending on the size of data loss and the reason behind it.
The match rate is relevant only for the product level data that is fundamental for product performance management. Campaign level data is downloaded a different way.
To learn more about maintaining 100% match rate, read here: How to Improve Match Rate.
Tracking Rate
Tracking rate shows how much spend from dynamic ads or campaigns is connected with revenue. If the tracking rate is 100%, it means that all dynamic campaigns and ads are tracked and can be connected with the revenue to confirm that some portion of the revenue came from dynamic ads.
This allows the correct product ROAS to be calculated, as the revenue from dynamic ads and the spend from dynamic ads.
The tracking rate is displayed as a single value, shown in the Market Health. A value of 100% indicates that either:
All ads are using automatic tracking
In the case of manual tracking usage, that all revenue has been sorted out as static or dynamic and is able to be counted
If the tracking rate is below 100%, it means that there is revenue which is unable to be categorized as from static or dynamic ads, and therefore must be manually assigned.
To learn more about maintaining 100% tracking rate, read here: How to Get 100% Tracking Rate.