This document discusses metadata and taxonomy. It begins by defining metadata as "data about data" and provides examples of metadata like an artist, genre and year for a song. It then discusses taxonomy and how it is used to classify items in a hierarchical manner, providing an example taxonomy for different types of animals. The document concludes by discussing how to build a taxonomy and apply metadata in SharePoint to improve findability of information.
2. Understanding Metadata
A supplement to: Practical SharePoint 2010
Information Architecture by Ruven Gotz
Buy the Book (or Kindle):
http://amzn.to/JnxlcC
3. Intro - Animated
@ruveng
spinsiders.com/ruveng
Toronto
ruveng@navantis.com
Ruven Gotz
40. Shared Drive Zoo
Taxonomy X:
Sales &
Production
Marketing
Marketing Sales Web Design Newsletter Social
Commercial Industrial Government Healthcare
Labs Hospitals Clinics Mobile
Private Public
Large Medium Small
Urban Rural
Not
University
Associated
43. Shared Drive Zoo X:
Sales &
Production
Marketing
Marketing Sales Web Design Newsletter Social
Major
Commercial Industrial Government Healthcare
Hospitals
Labs Hospitals Clinics Mobile Colleges
Private Public Big Small
Large Medium Small
Urban Rural
Not
University
Associated
54. A SharePoint Taxonomy (Metadata)
Customer Type Sector Size Location University
• Lab • Private • Large • Urban • Yes
• Hospital • Public • Medium • Rural • No
• Clinic • Small
• Mobile
61. Drug
Vacation Request
Reimbursement
Name _________ Name _________
Emp. # _________ Emp. # _________
Date _________ Date _________
Dates Requested: Drug Used:
From __________ Name __________
To: __________ Cost: $ _________
Manager ________ Manager ________
Approved Y/N Approved Y/N
73. Increasing the likelihood of metadata success
Compliance
Search & Discover
Records
Search
Management
Magic Metadata
Intersection Business
Process
Day-to-day workflow
77. Thanks for your attention
Buy the Book (or Kindle):
http://amzn.to/JnxlcC
Practical SharePoint 2010 Information
Architecture by Ruven Gotz
Notas do Editor
[ANIMATED]
Everyone knows…
It’s “Data about Data” as Einstein proved all those years ago
This answer helps exactly no-one.
I won’t tell you yet but…It is an iterative process – you won’t understand it right away, but you will circle in towards understanding over time[ANIMATED]
Metadata is a new concept for manyUse of metaphors to explain the concepts
Metadata is a new concept for manyUse of metaphors to explain the concepts
If you were to ask a child: What does a cow say? How would they answer?
Ok, so that was goofing around – now let’s get serious.
The music is the contentYou can know a lot of facts about the album:PrincePop/Rock1984You can know all the facts, but it doesn’t substitute for the content (the music)(Purple Rain example originally suggested by Yoav Lurie)
How do you sort CD’sArtist?Title?Year?Genre?Colour?You have to decide up-front – and stick to it –because the objects are physical
What if the store was full of unlabeled tin cans?You would need to open every can to see if had what you wanted(Tin can example originally suggested by Serge Tremblay)
Now we don’t need to open each can, but they are all in a jumble and you have to pick up each can to check if has what you want.
Items are grouped by area (canned fruit, canned sauce, canned vegetables)Signs point you to the correct area so that you can quickly find what you need.BUT: Because the objects are physical, you need to pick a method and stick to it
This uses the base metaphor that we live with every day.The concept of a “file” and a “file folder” as a way of storing digital data is a metaphor taken from the world of paper managementIt has become so ingrained, that we think of it as natural, but it’s not: It was invented in 1983 by Apple (wikipedia)
Old apple interface from the 80’s
All your files are stored in one folder and their names are completely meaninglessThis is like the unlabeled cans: You have to open each file to see what it contains
You have a bit of a better situationThe naming convention lets you find the file you need (but there’s no way to sort by year)Rely on users to follow the naming convention (religiously)
A ha!Now we’re in great shape. We’re like the supermarketStructured and LabelledBUT...
... then, you hire a summer internWho doesn’t know the folder hierarchy and makes up their own
Findability is challengingPutability is the real problemThis is Bill English’s word for knowing where to save a documentWhat if we could make putability easier while also improving findability?This is the promise of metadata
Data about dataYes, but not enough info Seth Maislin of Earley & Assoc. says it's the "Is-ness" of something:This 'is' a contract. That 'is' a pop album.For us it enables findability, policy and processFindability for locating the right documentsPolicy – records managementProcess – Status of a business process (e.g. Not started, In process, Complete, Approved, Archived)
So, let’s create an alternative structure that is logically equivalent, but that makes putability much easier while preserving findabilityBy the way: One way to start to figure out an organization’s metadata is to look at the folder names.You will probably not want to simply copy this, but it can be a good guide/starting point
It’s not this… (visual joke)[ANIMATED]
It’s not this… (visual joke)[ANIMATED]
It’s this…Not really this, but let’s use these creatures to understand.
Swedish botanist, physician, and zoologist, who laid the foundations for the modern scheme of binomial nomenclature. He is known as the father of modern taxonomy
This works because it’s really about governance – this is a stable structure that can’t be changed by just anybody:Changing this structure requires a world-wide meeting of the top scientists in the field, usually involving name-calling and fist-fights (or so I’ve heard)[ANIMATED]
Did you catch the subtle change here.The taxonomy is now of your ‘X’ drive.
And this is a common result
The asymmetry is that you’ll spend less time looking for a place to save something then you will looking for something after the fact.
Problem here is lack of governance – anyone can add any folder anywhere anytimeThis boils down to the ‘putability’ problem – I’ll search for a long time to find a doc, but not for long to see where to put it.[ANIMATED]
Once you’ve migrated your x drive to SharePoint, and all the promised benefits fail to emerge, The reaction is: (next slide)
Never, ever, use folders!Except when it makes sense to do so.[ANIMATED]
Never, ever, use folders!Except when it makes sense to do so.[ANIMATED]
[ANIMATED] Here is one reason to use folders: Application of security and then simplifying it for the user by using the ‘no folders’ view.
I’m not Carl, but let’s talk about why this works.After all, it’s the same as a directory treeThe difference is governance
Is this too many to ask for?Do we force users to answer all these questions/enter all this data?
Instead of confusing people with the SharePoint interface, I use a familiar tool: ExcelUsing some simple macros, I am able to illustrate the power of filters and views.There’s no free lunch however: People now have to enter metadata.We can simplify this by defaulting values like “Date” to today and “Year” to current year.We can leverage content types as well
Explain metadata and then use this worksheet for ‘homework’
Think of them as different forms with slots to fill in.Two documents may have overlapping slots (or, metadata).It may make sense to store these two types of docs in the same library (HR Requests), but use content types to drive workflow, policy and prompt users only for the metadata that applies.[ANIMATED]
Think of them as different forms with slots to fill in.Two documents may have overlapping slots (or, metadata).It may make sense to store these two types of docs in the same library (HR Requests), but use content types to drive workflow, policy and prompt users only for the metadata that applies.[ANIMATED]
[ANIMATED]
Using mind-mapping tools to build the taxonomy from the homeworkI use MindJet MindManager – and I like and highly recommend it.There are other tools that are less expensive.
Search – Find docs: eg. Subject, Status, Pub type, Author, YearRM – manage compliance: Retention period, confidentiality status, FOI rulesProcess – Day to Day: Status, assigned to, Due Date, etc.
Pragmatic & Outcomes focused
The answers are not clear-cut, everyone has to be on the same page.These decisions are not made by the consultants, or even IT – it has to involve the business
Find the common denominatorShoot for the magic subset