Same and differentarchitectures for mass-uniqueness
Tom Graves, Tetradian ConsultingOpen Group London, October 2013
the futures of business
Hi.
(let’s not bother with the PR-stuff?)
It begins with a pin, perhaps…
CC-BY Creativity103 via Flickr
…or maybe some dust?
CC-BY storebrukkebruse via Flickr
What isenterprise-architecture?
But perhaps better start witha much-argued question:
Maybe more to the point –
why enterprise-architecture?
– it’s because
things work betterwhen they work together,
on purpose
Why enterprise-architecture?
Why enterprise-architecture?…which implies further questions:
•Things – what things? and who decides?
•Work – ‘work’ in what sense? what work?
•Better – ‘better’ for what? or who? who decides?
•Together – what kind of ‘together’? how? why?
•On purpose – who chooses the purpose? for what?
One place to look for clues is in the enterprise’s balance between same and different…
Into practice…
(Each ‘Into practice…’ sectionprovides a brief moment to exploreimplications of the preceding ideas.
Use the text on the slide to guidea quick review of related design-themes
in your business-context.)
“What’s the story?”A quest for productivity
Let’s go back to an earlier time…
there’s not much technology…
maybe some of it a bit strange…
everything made by hand…
everything different, unique…
almost nothing standardised…
until this guy, in this book…
looked at how pins were made…
CC-BY Creativity103 via Flickr
applied it to other industries…
via sameness…
CC-BY toktokkie via Flickr
more sameness…
a lot more sameness…
then applied to work itself…Richard Arkwright’s Cromford Mill
CC-BY-SA CaptainScarlet via Wikimedia
mass-sameness…
mass-sameness…
CC-BY aleutia via Flickr
mass-sameness…
CC-BY Vlima.com via Flickr
mass-sameness everywhere.
CC-BY-SA MysteryBee via Flickr
the real value of samenessis that it’s easy to scale
and easy to make efficient- creating huge productivity
Why sameness?
But there’s a catch…
the differencesnever went away…
and if we pretendthat everythingis sameness…
we court disaster…
Into practice…
In what ways do your systemsdepend on sameness?
What would happenif the sameness wasn’t there
to depend on?
What happens with anythingthat won’t fit those expectations?
“What’s the story?”There’s always an exception…
What name in your system?
Typical UK-style name-structure for database:
•Title (mandatory: select from picklist)
•Forename (mandatory: 30 characters max)
•Middle-name (optional: 30 characters max)
•Surname (mandatory: 30 characters max)
•Suffix (optional: select from picklist)
Easy, right? – well, let’s take a real example…
What name in your system?
UK-style name:
•Mr Pablo Diego Ruiz
What name in your system?
UK-style name:
•Mr Pablo Diego Ruiz
Full legal birth-name:
•Pablo Diego José Francisco de Paula Juan Nepomuceno María de los Remedios Cipriano de la Santísima Trinidad Ruiz y Picasso
What name in your system?
UK-style name:
•Mr Pablo Diego Ruiz
Full legal birth-name:
•Pablo Diego José Francisco de Paula Juan Nepomuceno María de los Remedios Cipriano de la Santísima Trinidad Ruiz y Picasso
You probably know him as:
Driver’s licence, please?
Real simple, right?
Hmm… maybe not so simpleafter all?
The same for everyone, surely?
Hensel twins’ driver-licences >>
Driver’s licence, please?On the flight: One ticket, one seat, two passengers, two passports
In the car: Two drivers behind the wheel, each legally liable
What keeps executives awake at night?
And here’s a real casefrom my own consultancy-work
in enterprise-architecture…
Executive #1: PR disastersGovernment department
(in social-work sector)
Real (if unofficial) business metric:
Number of daysbetween bad headlines
in the newspaper
Executive #1: PR disasters
Real newspaper headline:
Department fails again!
Ten life-critical incidentsin just one suburbstill not resolvedin 2½ months!
Answer: architecture of the enterprise
What actually happened?
Incident
for one incidentReport
Report
Report
ReportReport Report
Report
Report
Report
Report
Ten incident-reports
Key-field: Date of Birth(for unborn child…)
Executive #1: PR disasters
Moral of this story:
Every automated systemneeds an option for manual override
Into practice…
What exceptions are therethat could break
your current systems?
How do you find out about thembefore they break your systems?
What workarounds would you needto keep your systems going?
“What’s the story?”A bit of theoryon uniqueness…
…dust is everywhere…
CC-BY storebrukkebruse via Flickr
‘Cantor Dust’
Start with everything-different
‘Cantor Dust’
Find an area of samenesswithin all of that uniqueness
‘Cantor Dust’
Within each remaining block of difference,find another region of sameness
‘Cantor Dust’
Within each remaining block of difference,find another region of sameness
‘Cantor Dust’
Within each remaining block of difference,find another region of sameness
‘Cantor Dust’
Within each remaining block of difference,find another region of sameness
‘Cantor Dust’
Repeat the same processall the way to infinity…
no matter how far down we go,there will always be uniqueness…
…and every one of thoseapparent ‘samenesses’ we found
is also different from every other…
- uniqueness in the sameness…
Into practice…
How much at present do you designagainst uniqueness?
If uniqueness is a fact of nature,is trying to design against it
even a viable option?
Should you design for uniqueness?If so, how?
“What’s the story?”Uniqueness– how and why
Mass-uniquenessis uniqueness at scale
– where differenceor uniqueness
is a central factof the work itself
What is mass-uniqueness?
Mass-uniquenessSome everyday examples:
•Healthcare – unique needs, Hickam’s Dictum
•Customer-service – unique needs, ‘long-tail’
•Clothing – fashion, body-types, shapes, sizes
•Education – every student is different, unique needs
•Information-search – unstructured, natural-language
•Farming – weather, micro-climate, soil-types
•City-planning – topography, geography, particularity
standardised
A spectrum of uniqueness
customised unique
standardised…
CC-BY-NC-ND actiononarmedviolence via Flickr
customised…CC-BY-NC-SA Doctress Neutopia via Flickr
customised…CC-BY-NC-SA Doctress Neutopia via Flickr
unique…© Courtesy of 3D Systems
unique…
© Courtesy of 3D Systems
uniqueness…
© Courtesy of 3D Systems
uniqueness makes it something to celebrate…
© Courtesy of 3D Systems
Into practice…
How much mass-uniquenessexists in your business-context?
How much do you already designfor that uniqueness?
How do you supportthe required uniqueness at scale?
“What’s the story?”A question of perspective
Perspectives and journeys
Service-delivery is a journey of interactionswhere ‘inside-out’ (the organisation’s perspective)
touches ‘outside-in’ (the customer’s / supplier’s perspective)
Outside-in…
CC-BY Fretro via Flickr
“Customers do not appear
in our processes,we appear in
their experiences.”Chris Potts, recrEAtion, Technics, 2010
A stakeholder is anyonewho can wielda sharp-pointed stakein your direction…
CC-BY-NC-SA evilpeacock via Flickr
Stakeholders in the enterprise
(Hint: there are a lot more of them than you might at first think…)
Narrative and storyhelp us to identify
the exceptionsand uniquenesses…
The role of narrative
Technology
CC-BY-SA xdxd_vs_xdxd via Flickr
Process
People
The usual EA view
Stage
CC-BY-SA xdxd_vs_xdxd via Flickr
Scene
Actor
ActorStage
Stage
Stage
A narrative-oriented viewStage
Scene
Scene
Stage
Into practice…
What changesas you shift the perspective
from inside-out to outside-in?
What do the narratives tell youabout uniqueness in your business?
What do you need to changeto make best use of this?
“What’s the story?”SCAN – making senseof uniqueness
“Let’s do a quick SCAN of this…”
Order and unorder
CC-BY bobaliciouslondon via Flickr
“We have a rule for everything!”
CC-BY bobaliciouslondon via Flickr
Hmm… let’s do a quick SCAN of this…
“Insanityis doingthe same thingand expectingdifferent results”
(Albert Einstein)
ORDER(IT-type rules do work here)
Take control! Impose order!
“Insanityis doingthe same thingand expectingdifferent results”
(Albert Einstein)
“Insanityis doingthe same thingand expectingthe same results”
(not Albert Einstein)
ORDER(IT-type rules do work here)
UNORDER(IT-type rules don’t work here)
Order and unorder
A quest for certainty: analysis, algorithms, identicality, efficiency, business-rule engines, executable models, Six Sigma...
SAMENESS(IT-systems do work
well here)
UNIQUENESS(IT-systems don’t work
well here)
Same and different
An acceptance of uncertainty: experiment, patterns, probabilities, ‘design-thinking’, unstructured process...
THEORY
What we plan to do, in the expected conditions
What we actually do, in the actual conditions
PRACTICE
Theory and practice
algorithm guideline
rule principle
Different types of decision-guides apply in each ‘domain’
Sensemaking guides decisions
Guidelines for design
order unorder
fail-safe(high-certainty)
safe-fail(low-certainty)
plan
actual
Waterfall(‘controlled’ change)
Agile(iterative change)
analysis(knowable result)
experiment(unknowable result)
Why we need people…
What is always going to beuncertain or unique?
(‘Messy’ – politics, management, wicked-problems, ‘should’ vs ‘is’, etc.)
What will always be ‘messy’?
Wherever these occur,we’re going to need human skill…
Machines and people
order(rules do work here)
unorder(rules don’t work here)
fail-safe(high-certainty)
safe-fail(low-certainty)
analysis(knowable result)
experiment(unknowable result)
MACHINES PEOPLE
Waterfall(‘controlled’ change)
Agile(iterative change)
misapplied difference - ‘special cases’ -
creates inefficiency
misapplied samenesscreates failure-demand– a key cause of waste…
Into practice…
Trying to apply rules to everything,or to automate everything,
will cause your system to fail.
How do you identify the right balancebetween sameness and difference?
How will you avoid inefficiency,or failure-demand?
“What’s the story?”Balancing samenessand uniqueness
Taylorist-type modelstend to assume that everything
is a machine to ‘control’…
Find the right fit!
people will oftenrelate to machines
as if they’re other people…
Wrong and right…
order(rules do work here)
unorder(rules don’t work here)
PEOPLEas MACHINES
PEOPLEas PEOPLE
CC-BY justin pickard via Flickr CC-BY andré luís via Flickr
Right and wrong…
order(rules do work here)
unorder(rules don’t work here)
MACHINESas MACHINES
MACHINESas PEOPLE
CC-BY-SA izzard via FlickrCC-BY-SA MysteryBee via Flickr
How we really think…
CC-BY Brett Jordan via Flickr
Use context-maps such as SCANto identify
what may or must changewhat is or is not certain
how these vary over timeand what to do with each
Mapping the context-space
A surgical example…
patient identity
surgery plan
emergency action
theatrebooking
consumables
pre-op complications
family behaviour
surgical-staff availability
change oftheatre-availability
action-records
equipmentplan
patientcondition
verify identity
NOW!
before
certain
uncertain
A surgical example…
patient identity
theatrebooking
consumables
action-records
equipmentplan
verify identity
we need to be certain about all of these
NOW!
before
certain
uncertain
A surgical example…
surgery plan
surgical-staff availability
change oftheatre-availability
patientcondition
we expect(and plan for)
uncertaintyabout these
NOW!
before
certain
uncertain
A surgical example…
emergency action
pre-op complications
family behaviour
we don’t expectthese to happen,
but we need contingency-plans
and guiding-principlesfor all of them
NOW!
before
certain
uncertain
Into practice…
How would you map the right fitfor each type of context?
How would you ensure you don’ttreat people as machinesor machines as people?
How will you managethe inherent uncertainties?
“What’s the story?”Uniqueness, changeand governance
standardised
Balancing the spectra…
customised unique
sameness uniqueness
high-probability low-probability
high-dependency low-dependency
reusability bespoke
low rate of change high rate of change
We need architecturesthat express that balance
between sameness and uniqueness,and other trade-offs across the space…
Architectures and governance
…and governanceto guide relative-positioning
and changes over timebetween backbone and edge
Architectures for change
BACKBONE EDGE
Into practice…
What do you need, to balancesameness and differencecertainty and uncertainty
across your whole business-context?
What architecturesdo you need for this?
What governance do you needto manage their changes over time?
Same and different
Some key take-aways, I hope?
•Many industries depend on mass-uniqueness
•Sameness and efficiency are important, but over-focus on sameness can fail, lethally
•Uniqueness is inherent and unavoidable
•Need ‘just enough sameness’ to support scale
•Work with uniqueness, not against it
“What’s the story?”Thank you!
Contact: Tom Graves
Company: Tetradian Consulting
Email: [email protected]
Twitter: @tetradian ( http://twitter.com/tetradian )
Weblog: http://weblog.tetradian.com
Slidedecks: http://www.slideshare.net/tetradian
Publications: http://tetradianbooks.com
Books: • The enterprise as story: the role of narrative in enterprise-architecture (2012)
• Mapping the enterprise: modelling the enterprise as services with the Enterprise Canvas (2010)
• Everyday enterprise-architecture: sensemaking, strategy, structures and solutions (2010)
• Doing enterprise-architecture: process and practice in the real enterprise (2009)
Image-credits: Slides 64-67 courtesy of 3D Systems: http://www.bespokeinnovations.com/ Other photo-images via Flickr or Wikimedia, as shown on each slide
Further information: