20191011 - LASCOT2019 - Three Cups of Tea, Jabe Bloom @cyetain

by Thierry de Pauw on


3 cups of tea with a map: having conversations with Wardley maps, Jabe Bloom @cyetain


PhD in Transition Design
I don't research on people, but with people:
=> Practice-Based Research aka Action Research: doing research by doing it with people

the way for change to happen is to have 3 conversations

the way we are going to do this today is to have conversations initiated with maps.

I love Simon, but I think most of the work of Simon is towards C-level
=> difficult for teams
Lots of what we are going to do today are this team conversations allowing these teams to then have these conversations with C-level

Two types of logics:
- Propositional logic: prove what is true
- Inquiry logic: the logic I follow is to make questions occur
these two lead into each other

maps are helping with the inquiry (I think ???).

what ends up on your map is what is being discussed

  • @lunivore: “When you map, the maps constrain the questions you can ask. The maps aren’t there to get you to a valid answer. They’re there to get you to a valid proposition that you can test.” @cyetain #WardleyMapping #lascot

Intro to Wardley Maps:

At top, you have the visible things: user needs
Underneath: the invisible things necessary for the user needs

Evolutionary axis: Genesis, Custom built, Product, Commodity
- @lunivore: “Evolution happens in an ecology. One of the things I see is that people just redraw their architectural stacks, not an ecology; and I think that hides things which are important.” @cyetain #WardleyMapping #lascot

Ecology of Components:
- for any component in a marketplace, there will be competition
- this means: similar components will attempt to provide for the same needs as other components by either differentiating or providing operational efficiencies

Value Chain:
=> outcomes: the value provided by using the data (output of IT)

  • @lunivore: “Outputs are usually quantitive; they can be counted. Bricks, houses are outputs. A stable home is an outcome. Outcomes are richer; qualitative; require evaluation.” @cyetain #WardleyMapping - and he’s just put up a map with “bowl of oatmeal” and “satiation”, awesome! #lascot

Simon got value chains from Porter

A company's value chain is a system of interdependent activities, which are connected by linkages. -- Porter, How information gives you a competitive advantage

The Tea map: bottom - warm water, kettle and power => can be used to produce something else than producing tea

The bottom of the Wardley map is stable regimes of production.

Please note: if you work with Architects and you take an architecture stack and put it on a Wardley map, you may miss the ecology

  • @tdpauw: RT @lunivore: “Evolution happens in an ecology. One of the things I see is that people just redraw their architectural stacks, not an ecology; and I think that hides things which are important.” @cyetain #WardleyMapping #lascot

  • @lunivore: “The options available to you, they’re part of the ecology. So I want to point at ‘Oracle’ and say, what other providers are there? What other components could do that job?” @cyetain on difference between Architecture Diagram & Map components #WardleyMapping #lascot

Tragedy of Commons

when you have a non-competitive situation in place => people will manage the commons
the minute someone introduces the game theory => commons get destroyed

In a lot of businesses, there are common things: things you share
Example: physical infrastructure you work in

This problem can only be applied to things you consume. If you don't consume it, the common is not a problem.

The key is the idea of the management of consumables
- some things are consumable: examples functions and data
if more people use the login-function, does that function gets used up?
- some things increase in value when they are consumed

  • @cacorriere: Functions & data are not consumables in an IT org, meaning their value increases with use instead of decreasing with use. @cyetain #lascot

Differentiation, Disruption and Far-From equilibrium
can you think of some examples of differentiation? example: the new iPhone camera

certain things are valuable because they differentiate, some things get value because we use them and some things get value because we use them better (efficiency). @cyetain

There is a difference between a Backlog and a Portfolio of Options! @cyetain
=> story mapping + experiments to validate your options

In IT there are general 3 economic frameworks

Genesis -> Custom -> Product -> Commodity
differentiation scope scale

  • the economy of scale, the economy of efficiency: the predominant economic framework, mathematically proven, generally accepted => hard to go against this
  • the economy of value, differentiation

these 2 economic philosophies hit each other and can't work together:
- differentiation will not be cheap, scale will be cheaper

in the middle you have scope:
- you do have to differentiate and you do have to be efficient

difference place tend to be product teams and UX teams => the value they provide is as thin as possible because they know

scope place is the platform teams: only measured by functions used by the difference teams - if the function is not used it has any value

scale place ...

  • @lunivore: “The economies of difference, scope and scale... scope’s the one that doesn’t get talked about. There, measure value by speed of adoption. You put a new function in your platform; how quickly do the differentiation teams use it?” #cyetain on winning plays #WardleyMapping #lascot

In Cynefin terms: difference is complex and scale is complicated


  • @lunivore: “The economies of difference, scope and scale... scope’s the one that doesn’t get talked about. There, measure value by speed of adoption. You put a new function in your platform; how quickly do the differentiation teams use it?” #cyetain on winning plays #WardleyMapping #lascot

  • differentiator:

    • I want you to make disposable things, it does not have to be tested, you use Ruby on Rails
    • only when the option has been validated, you start creating production code with a little tests
    • when it starts to be adopted, you add all the rest

==> user needs always change

  • scope: you want to accelerate the use of the platform => you need to make it so people want to use it
    • instead of finding the market, you need to capture the market as long as top line moves up faster than bottom line

=> our platform == options: we need to recommon our functions and data, we need to minimise the cost of change

  • @lunivore: “Scope... I’m going to measure you on how many people use your platform; but you can’t force them to use it. You have to make them want to use it.” @cyetain #WardleyMapping #lascot

    • scale: aggressively reduce the bottom line => MLREAM: Moore's Law rules everything around me, the economics of hardware and infrastructure are at a scale FAR BEYOND our control

Economy of Differentiation:
- defined by the economoic advantage of creating somehing novel in the market
- high failure rate

Economy of Scale:
- driven by reproducing consumable good with ever more efficient processes
- no standard no kaizen
- shrinking margins means growth is driven by market consolidation: when commodity, there are few competitors
- small players can't achieve the same scale as market consolidators ... so have limited growth ... and limited options

Economy of Scope:
- driven by leveraging good that are not consumed in use
- they are Ecological:
- they serve differentiators by accelerating creation
- they serve scale players by reducing variation

Platform people should absolutely not build a platform. They should watch the differentiator teams and discover what they are using. Hey, you are using a login function and you too. Ok, we are going to create a login function. @cyetain

Data Gravity

Dave McCrory

Functions are ephemeral, they can move everywhere when they want.
Data are not ephemeral, they have gravity.

you can move your functions quickly but if they depend on data, you can't move your functions rapidly.

Data Mapping

  • data traversals: where should it be valuable
  • data wells: where should it be located

data traversals:
- genesis - custom built - product - commodity
- un-modelled - divergent - convergent - modelled

data wells:
- genesis - custom-built - product - commodity
- local - distributed - colocated (schema on write) - consolidated (schema on read)

critical: latency and throughput
we tend to not measure latency and throughput when data is local

Functional Ephemerality is defined by: cost to spin it up + execution + cost to tear it down
most orgs are not aware of that, they tend to spin up the function and never tear it down

you have to be careful when saying: we are going to the cloud and our costs will go down. Well, that function is executed 30 times a day. How much time does it take to spin it up?

Trying to find one economic framework for prioritising backlogs is not feasible. You need to use all three: differentiation, scope and scale @cyetain

When there is no compelling need for Movement ... (when you are dominating a stable market)
Focus all effort on reducing the cost of change.