Purview: we have a Problem
"My great colleague Andreas Bruun Str?bek wrote an important blog post that gives some important insights into why Purview Implements fail. (The link is in Danish.) I thought it was so important that I translated it.
How to get success with Purview
Many organisations are trying to use Microsoft Purivew to clean up their illegal privacy data but often end up throwing in the towel.
However, there's hope! The main issues with Purview can be summarized briefly:
You need to build and maintain a very complex search engine that nonetheless has a high error rate and is incredibly slow at searching through data. Any mistakes or changes mean starting over. Not all relevant data can be handled by Purview, and when it finally comes to deleting data, you're essentially deleting blindly without knowing if Purview has made an error.
These issues are often dealbreakers for most organizations that want to control their own data. I've thus briefly expanded on what Purview struggles with and what it can achieve despite these challenges.
So where does Purview falter?
Poor Data Identification
Purview offers a few predefined searches, which are far from adequate and are not very precise. You must establish the thousands of types of searches necessary for compliance with laws like GDPR and soon NIS2. This is not necessarily a dealbreaker, but it is a massive task requiring a dedicated team of linguists and Microsoft consultants. For each language used in your organization, a separate setup must also be created. Care must be taken, for example, to ensure that a German sequence number is not confused with the Danish social security number. The complexity of ensuring proper searches is both a huge task and critical for success.
Very Slow Scanning
Purview is an extremely slow data scanner, so you might find yourself waiting years just to get your data scanned. For example, if you have 250 employees each with about 100,000 data sets (emails, attachments, files, etc.), it would take approximately 365 days to scan all the data (calculate for yourself if there are more of you). And that's just one scan. Since you can only run a scan lasting 7 days at a time, the project process of completing the scan of all data is also cumbersome.
Re-scanning Often Necessary
You'll naturally need to adjust your searches as you learn more about your data, what you want to scan for, and if you wish to minimize false positives. Each change requires re-scanning all the organization's data.
Labeling - But Only on Selected Data
Labeling is a key functionality in Purview. It’s crucial that the right data are marked with labels, so they can subsequently be managed by Purview. However, only a very select few data types can be scanned and labeled by Purview. The rest of the organization's data must still remain uncontrolled.
Blind Cleanup
Few dare to tidy up blindly without knowing exactly what they are deleting (data often wrongly selected). But this is what happens in Purview. The employee/data owner is not involved, and instead, a decision to initiate deletion is made from a central position. This can have enormous business consequences when essential data disappears. Who dares to press the delete button?
But as mentioned, there is a way to keep Purview, get it started, and succeed with it:
Danish Data & More’s Solution
Data & More's solution is specifically created to handle all these issues where Purview fails. You might use the solution from Data & More alone, or as the solution that handles the heavy lifting, allowing your organization to quickly start using Purview for what it does best.
领英推荐
About Data & More’s Solution:
The solution is ready to use. There’s no need for configuration or programming.
It contains hundreds of thousands of predefined and thoroughly tested searches in many languages.
The solution scans quickly and can handle huge amounts of data (petabytes).
If the search criteria are changed, it immediately affects all data - there’s no need for re-scanning.
The solution always keeps track of all data across all connected data sources - including new data.
Labels can be set on all types of data - which can then be used by Purview, DLP, etc.
The employee is automatically involved when necessary.
The employee can double-check what has been marked for deletion before it is deleted.
Selected data can be found in seconds (e.g., for access requests) across all data.
More About Data & More’s Solution
Fundamentally, Data & More's solution does the hard work – quickly and efficiently scanning all data, applying the right sensitivity labels to the correct data in all relevant data sources, and involving the employee if data also needs to be deleted. Data & More has developed and continuously maintains a classification for GDPR, CCPA, PIPEPA, etc., based on billions of data sets (emails, attachments, files, etc.), which can be used directly in Purview.
Learn more here: [Data & More](https://dataandmore.com/).
Thus, Purview is given the optimal conditions for success.
Bonus Info:
Data & More will be presenting on this topic on May 14th. This might be of interest to you: AI Holds the Keys to Your Business
This post is based on the following sources: