When is a catalog not a catalog?

I blogged not too long ago in regards to the excessive stage of hype and confusion throughout Data and Analytics simply a few months in the past.  Here is the unique weblog from March 2023: Summing Up Three Days at Gartner’s Data and Analytics Conference in Orlando, Florida, USA.  The hype and confusion I registered within the US was additionally seen after I attended our D&A Conference in London, UK, a few weeks later.  Interestingly the hype and confusion was markedly decrease after I was in China simply a few weeks after that.

One of the factors of confusion is with catalogs – or knowledge catalogs – or analytics catalogs or metrics shops.  The proven fact that there are completely different names is one factor.  Here I repeat what I wrote within the authentic weblog:

Use instances for a knowledge catalog

Analytics use instances are fairly completely different to governance use instances.  Too usually they’re conflated.  Several 1-1s requested the way to get enterprise of us concerned and excited to work with a knowledge catalog in assist of a governance program.  That is the incorrect query.  For governance, enterprise is already however they’d solely actually care about what ought to be named a glossary.  Even the information dictionary is perhaps used selectively by a steward (within the enterprise) for root trigger evaluation.  Those are subsets of a a lot bigger catalog.  See Quick Answer: What Are Differences Between a Data Dictionary, Business Glossary and Data Catalog?

I used to be on a vendor briefing at this time with a vendor (who will probably be anonymous) the place they described what is of their “knowledge catalog”.  After I listened and probed, it appears there is a lot much less “catalog” within the catalog!  Here is what the seller stated, and I paraphrase:  “The catalog is the place the place we stock your knowledge.  We additionally retailer all of the historical past, lineage, insurance policies, knowledge homeowners, guidelines, situations, and supporting knowledge to assist with governance.”  This is not a lot of catalog in any respect, and is nearer to being a D&A governance answer and stewardship answer.

To catalog, or not to catalog

As I famous above, there is a clear use-case for a catalog in analytics.  When a enterprise analyst or knowledge science chief is constructing a mannequin, they’ll usually begin out by looking a catalog for knowledge. This could embrace a seek for knowledge units and even earlier analytics fashions that is perhaps leveraged.  These catalogs might be known as knowledge catalogs (for knowledge or knowledge units), or metrics shops (for fashions and metrics).  Why some a part of the market began to make use of “metrics retailer” and one other half went with “knowledge catalog” or “analytics catalog” is past me.  They are all catalogs.  They are inventories of issues.

What this vendor known as “a knowledge catalog for knowledge governance” throughout at this time’s briefing was actually talking and promoting to the D&A governance use case.  And each use instances, analytics and governance, are being offered a catalog of some type.  This is the supply of confusion.  The governance use case is nothing just like the analytics use case.  No enterprise function of their rightful thoughts sits there throughout a regular day and asks themselves, “I ponder what knowledge to have a look at, at this time?”  The world of governance is exception primarily based.  It is not about creating metrics and fashions for evaluation – although there are some metrics and evaluation that must be completed.  That is complicated the purpose.

Being clear in regards to the use case

The work we do in D&A governance and stewardship (setting and implementing coverage) ought to not happen “inside a catalog” in any respect however “inside a answer designed to serve the wants of governance and stewardship”.  You would not use an EDW if you want ERP, nor would you discuss in regards to the database that sits on the foot of the ERP utility; its included.  Stewardship and governance options are what we ought to be referring too, not catalogs.  Even if a knowledge catalog performs a function deep contained in the physique of the governance and stewardship answer, it is not a catalog of any type.  ERP is not a database, although it accommodates one.  We ought to break up out and make clear the roles and use-cases for which expertise is being thrown.

One final level: Back to the use instances.  I’m taking slightly too many calls from shoppers who’re “turning off” their catalogs and marketplaces.  In the case of the catalog state of affairs is that the incorrect catalog was sole.  Perhaps a catalog that was initially deigned for analytics has been augmented with some governance functionality, and offered into that use case.  Some time later the top person realizes there is a hole between what they actually need, and what they’ve.  The reverse is additionally true too: promoting a governance answer, that occurs to incorporate a catalog, into the analytics use case.  That consumer additionally will get sad about 5 months into the mission.

Ask what you are able to do for the person, not what the person can do for you

Business roles will not govern knowledge in a catalog – they don’t have any use, no want, and its simply foolish.  They WOULD have some alternative to manipulate the glossary!  And guess what – there are different names that now we have collectively invented over time that are likely to imply the identical factor:

• Glossary
• Business metadata, and
• Master knowledge.

Those are the issues enterprise enterprise of us would care about, since they use that knowledge day by day.  The different 96% of knowledge within the catalog has little that means to them.  And for probably the most half, by no means will.

Even extra confusion

I noticed a slide through the vendor briefing and it is fairly per normal messages available in the market. The slide contrasted privateness purposes to safety purposes to governance purposes.  This is inconsistent.  Privacy and safety purposes are examples of governance apps.  More exactly they’re all purposes that deal with completely different governance coverage courses.  What was documented below governance apps by this vendor was actually a mixture of different varieties.  One or two have been coverage centered options themselves, equivalent to knowledge high quality.

There was additionally reference to stewardship options that ought to be used throughout all coverage areas, together with privateness and safety.  In impact there ought to be options that deal with a vary of coverage courses equivalent to safety/entry, privateness, high quality, retention, ethics and requirements.  See Effective Data and Analytics Governance Includes a Range of Policy Types.

Back to the longer term

After scripting this weblog, I requested myself: Surely I had written this weblog earlier than, sure?  Yes, I had.  Here is is: When is a Catalog No Longer a Catalog?    This was from 2018.  If you may afford the time, it should make you smile.  Its the identical message!

 

 

The put up When is a catalog not a catalog? appeared first on Andrew White.