High-level IT Blueprint

What does a project on a high-level operating model need to say about IT?  And what does it need to deliver to the IT architecture team?  This is the focus of this blog.   The detailed work on IT architecture and an operating model for the IT function will be done by the head of IT with support from an Enterprise Architect.  So the work that is done on a higher level operating model must take into account what the head of IT and his or her enterprise architect need.  I suggest the following

  1. The head of IT will need information about the strategy and the broader operating model that IT is supporting. This, of course, is the reason for creating an operating model in the first place.  It translates strategy into a high-level design for the operations and the organization, that can then be handed down to each function and each business unit, who will then do the next level of design.
  2. The head of IT needs to know who the business owner is for each important application.  With this information, the head of IT will know who to interact with, who to deliver value to and, hence, who are the prime customers of the IT architecture that will be created.
  3. The head of IT needs to know which of the software applications need to be part of an integrated enterprise system and which can standalone.  Since it is only IT that can mastermind the integration, it is important that IT knows which applications need to be integrated.  In this context, integration means that the applications can communicate seamlessly, that the data architecture and coding is the same in each application, etc.   Those applications that do not need to be integrated may not even need to be included in the remit of IT.  It is possible that the business owner can be responsible for sourcing, maintaining and running these applications, in the same way that employees are often responsible for their mobile telephones and the applications they have on them.  For example, an engineering team may have a CAD/CAM application supporting their design work which is sourced and maintained by the engineering team.
  4. The head of IT needs to know which software applications will need to be bespoke and whether IT will be expected to develop or commission these bespoke solutions.   Typically, the only applications that need to be bespoke are those that help create a competitive advantage for the company.  Hence, it is the operating model team, supported by the strategy people who are in the best place to identify where tailoring is likely to create advantage.  Everywhere else it is normally best to use a standard package and change the internal processes to fit the standard.  When dealing with business owners, IT needs to know whether to support or resist the special requirements that business owners are likely to request.

This list of four outputs from a high-level operating model project is all that needs to be done on information systems: everything else can be delegated to the IT steering committee – the next level of operating model design.  If you have a different view, please comment below.

Since I believe that operating model work should result in something visual – map, chart, table, etc – my co-authors and I (of the book Operating Model Canvas) have developed a high-level IT blueprint to capture these four points.  Lots of readers of this blog will have their own views about what the term “IT blueprint” means – and please feel free to comment.  This is a “high-level” IT blueprint, and it has been developed without much attention being given to the other forms of IT blueprint.

ch3-p12-it-blueprint-2slide1

The high-level IT blueprint is a table with the main steps in the value delivery chain listed along the top axis and the main organizational units listed down the side axis.   As Mikel Gutierrez likes to say, “when you match the organization against the processes, something magical happens”.  The magical thing that happens is that you are able to visualize the following design choices.

First, you can identify which organizational unit is involved in each of the steps of the value delivery chain.  So, in the example, the step “design product” involves customers/distributors, sales and specials unit: to design a specials product, the specials unit who will do the design must know what the customer or distributor wants and what the sales team has committed to the customer.

Second, you can decide whether this step requires a software application (or sometimes more than one).  The “design product” step does require software support.

Third, you can decide which organizational unit is the business owner of the application: it may be a committee of all of the units; it may be one unit; or it may be possible to sub-divide the application into two or more sub-applications and have multiple business owners.  In this case, the specials unit was the business owner.

Fourth, you can decide which software applications need to be integrated into the enterprise system (the pink boxes in the chart).  Typically these are the applications that touch multiple parts of the organization.  The software for designing specials did not need to be integrated into the enterprise system.

Fifth, based on your understanding of where, in the value delivery chain, competitive advantage is created, you can identify which applications are likely to need to be bespoke. The design application for specials was expected to be a source of advantage, hence it would need to be bespoke.

And, when you have made all of these choices, and recorded them on the table, you have produced  a high-level IT blueprint: a single page that delivers to IT the information that IT needs about the IT implications of the operating model.   Of course, you can rarely get it all on one page – mainly because you will typically have multiple value deliver chains.  However, you can normally produce a high-level IT blueprint with only a handful of pages.

Advertisements
Posted in IT Blueprint | Tagged , , , , , , | Leave a comment

Management systems and accountability

I have done work on the subject of accountability which has led me to the view that a person or unit is accountable when

“the stakeholders of that person/unit have the ability to reward the person for good work or create pain for the person/unit when the work is not good”.

Clearly this is best done directly by the stakeholder interacting with the person or unit, as for a business unit (if the customers are happy they buy more or unhappy they buy less; if the suppliers are happy they deliver on time, if unhappy they deliver late or deliver substandard service, etc). But accountability can also be achieved through the boss of the person or unit:  the stakeholder signals to the boss if things are going well or badly and the boss then administers the reward or pain.  This indirect mechanism is much less effective, because the boss often does not pass on the reward or pain, or amplifies it inappropriately or filters the signals based on the boss’s own views.

The wonder of accountable people/units is that they do not need much external managing.  The person or leader of the unit is plenty motivated to get it right.  Of course a mission for the person or unit that fits with strategy still needs to be agreed, performance still needs to be monitored, and, when the person or unit is failing, someone needs to be developed or changed.

So there is a tight link between management system – planning, budgeting, performance review, people review, risk management, continuous improvement, etc – and organization design.  If the organization has been structured into accountable units, the management system, particularly on the issue of performance review can be much lighter.

Before I continue, let me just back up and explain why I arrived at this view of accountability.   I run a research center at Ashridge Executive Education and we have some important stakeholders – executives who read our research, the larger school, member companies who give us money, participants on our courses, editors of journals, etc.   As I was thinking about accountability, I was recognizing that I feel very accountable and I was wondering why.  Also, I only meet my boss, the Dean, for about an hour every six months; so I was not feeling accountable because of the management system.  I realized that I was feeling accountable because each of these stakeholder has the power of reward or pain over me (I won’t go into the details – but it is because the research center is self-financing).

I also realized that success is about doing a sufficiently good job for each stakeholder so that the stakeholder remains engaged with the research center.  Of course one of the stakeholders is the mission stakeholder: for us executives of large organizations. For this stakeholder the objective is not “sufficiently good job” but something more ambitious.  For the other stakeholders we needed to think about the value we could give to keep them engaged – editors, course participants, member companies, central IT, etc.   It is much easier to solve this puzzle if the stakeholder has the ability to communicate directly and powerfully (through their behavior and choices).   So one of the difficult relationships has been with central IT.  They could create pain and reward for us.  But we had no way of delivering value to them so as to motivate them to reward us rather than create pain.

So what stimulated me to write this blog.  First, there are some important ideas here that should influence organization design and management system design.   Second, I came across this paragraph in a blog by Leon Tranter

“In Toyota car factories, this (the ability to ‘stop and fix’) is implemented via something called “Andon cords”. These are big cords that hang from the ceiling in the car factory. If there is a problem anywhere that needs attention, and cannot be quickly fixed by one person, they pull the cord. Pulling the cord shuts down the whole assembly line. Everyone stops working on whatever they are doing and comes over to help sort out the problem. This might sound insane, but it isn’t. This is Toyota we are talking about here, the most successful car business on the planet. They pretty much invented Lean and they know what they are dong. Andon cords encourage a culture where problems don’t lie around and take root. They get fix and fast. Which means the rest of the business can then restart and keep going.”

I don’t think the Andon cords actually stop the line, but they do signal to management and engineers that they should stop what they are doing and run to where the problem is.  This mechanism gives the worker the power to create pain for the engineers and managers.  It is this sort of mechanism that creates accountability.

In an article I wrote for McKinsey Quarterly, I suggested that one of the ways of stopping corporate headquarters in large companies from being a burden on business units would be to give managers in business units (maybe all managers) the equivalent of an Andon cord so that they could blow a “bureaucracy whistle” when activities by headquarters were time wasting or interfering and the “whistle”  would bring senior managers running to “fix the problem”.

Posted in Management system | Tagged , , | Leave a comment

Decision Grids

Decision Grids

An important tool in both operating model work and organisation design work is the decision grid.   The format is simple. Place the decisions you want to consider down the left hand side of a table and the organisation roles (like CEO, CFO, etc) that are likely to be involved in the decision along the top of the table. This gives you a grid of boxes where decision meets organisation role. In these boxes you then place a letter (or words) that describes the relationship that that organisation role has in the decision.

slide1

The best known mnemonic is RACI – standing for Responsible, Accountable, Consult, Inform.   But there are other mnemonics that I find more useful. So this article is about which mnemonic and how to get the best from the tool.

RACI

The RACI mnemonic was developed to help people think about activities rather than decisions, for example, the activity of producing the annual report or the activity of developing new customers.   In this context, Responsible means the person who is responsible for doing the activity: the project manager for the annual report or the business development manager for new business.   Accountable means the person whose job it is to make sure that the activity is done and done well: the CFO typically for the annual report and the head of business development for new business. Another way of thinking about Accountable is “who is the most senior person who will get fired if this activity does not happen or is done badly”.

There is a long tradition in academic literature that argues that responsibility and authority should be aligned. This means that the person who has authority should either also be responsible or should have the person who is responsible reporting to him or her.   RACI has been developed out of this tradition, with the word accountable standing in for the word authority. So it is for this reason that R and A are the first two letters of RACI. The last two letters standing for Consult and Inform do not need much explanation.   They enable those related to the activity, but with limited power over the activity, to be acknowledged.

slide3

The reason why I find RACI less useful than some alternatives is because the words accountable and responsible are often confused in people’s minds and do not have a clear, intuitive meaning. Also, the words are less useful for decisions than for activities.   My preferred mnemonic is RAPID

RAPID

RAPID was developed by consultants at Bain & Co in order to create a focus on decisions and decision processes. There is a famous Harvard Business Review article titled “Who has the D?” by Marcia Blenko (if I recall correctly).   The D stands for Decide: who is the person with the authority and power to make the decision? For example, who decides what the sales targets should be? Or who decides what text should be included in the Chairman’s statement in the annual report?   For the latter, presumably it is the Chairman. For the former, is it the head of sales or the head of finance or the CEO or …?  You can see from these examples that part of the skill of using this tool is to define the decision quite precisely: so it is possible to pin it on one of the people involved.

slide1

R stands for Recommend.  Often, in a decision process, the person who is responsible for doing something (the P for Perform) is also expected to make a recommendation about what decision should be made.  This recommendation is then approved (decided) or not by a boss or the leader of some committee.

I stands for Input: a person who is expected to have some input into the decision.   For example, the CEO might have some input to the Chairman’s statement, but not decide or recommend. The regional heads of sales and the financial controller might have some input into the sales target.

A is the most difficult letter in RAPID. It stands for Agree. It is mostly used where there are constraints on the decision such as legal or financial issues.   A person who holds an Agree role should sign off the decision based on whether it meets some standard, for example it does not break a law or require more money than is available.  In other words the remit of the person with an Agree role is quite limited.   Agree does not give the person the power to question the logic of the decision or recommend alternative points of view (but of course it does not stop someone offering their opinion.)  It just means that the person holding the D can ignore the opinion offered unless the opinion is limited to the specific standard or requirement (such as is it legal?).

slide2

One of the strengths of RAPID is the idea that there should be only one person holding the D. Hence the article title “Who holds the D?”

Other mnemonics

There are quite a range of other mnemonics in part because popular tools like RACI and RAPID both have weaknesses. Neither addresses the hierarchy of mangers or governance bodies above the decision maker, each of which typically has the power to overrule the decision maker.   Neither has letters for all the possible roles connected with an activity or a decision.

RASCI, standing for Responsible, Accountable, Support, Consult, Inform, is one alternative. RAS standing for Responsible, Approve, Support is another alternative.

slide4

Those working on operating models or organisation designs should not feel constrained by RACI or RAPID.   Be comfortable creating your own set of letters to address your particular situation, whenever one of the popular mnemonics does not address all the issues being raised or addresses more issues than it is useful to raise.

Posted in Decision grids | Tagged , , , | Leave a comment

Locations footprint

Location is one of the spaces on the Operating Model Canvas.  But I have not blogged about locations at any point.  This is partly because the topic is pretty self explanatory, and partly because I have found other parts of the operating model puzzle more interesting.   So let me say something about locations.

First, location is important.  Should you locate near customers or near suppliers or in a low cost country or where there are skilled employees or ….?  These are big operating model issues and the right answer is changing due to digital technologies.

But location is not just about which countries to be in.  It is also about which cities or towns in a country to locate in.  It is also about the type of buildings that are needed for the work and the people who do the work.   And it gets more detailed than that.  Location can be about factory floor plans.  It can be about which departments are on the top floor of an office building and which on the ground floor.  It can be about open plan or cubicles or partitions.  Mars is famous for its open plan even for the CEO and CFO.   Zara is famous for locating its Product Managers in the middle of an open plan space with the design teams located around the outside of the space.  This helps easy communication between designers and product managers, who know what is selling in the market.

So location has many dimensions.

Second, location is important because it is frequently connected with big spending: the land, the buildings, the equipment, etc.  When doing a business case for an operating model change, the locations costs are frequently a significant element.

Third, location is often a major constraint on operating model work because of legacy buildings and employees with deep location roots.  The best operating model is not always possible because of buildings or people.

Fourth there is little guidance on what should be included in a locations footprint chart.  Clearly it should visually communicate the locations through a map or a floor plan.  But what other information should be on the chart?  Here is a list of items to consider:

  • type of work carried out in the location
  • number of people in the location
  • logic for the location (which could be “legacy”)
  • information about the buildings and other assets at the location
  • important connections between locations

Please suggest some more items that you may have seen on a locations footprint.

Fifth there are few tools that I am aware of for helping you think about locations.  In different industries, there are industry specific tools.  So a supermarket company will have extensive tooling for identifying the best locations for its supermarkets.  However, it terms of generic tools, there are none.  As a result, I have developed a stakeholder tool for thinking about location (see exhibit).   This needs some work – but is a helpful starting point.  It is possible to take any activity or any asset and use the tool to consider whether there is a stakeholder logic for positioning that activity or asset in one location or another.

slide1

Posted in Locations | Tagged , , , | Leave a comment

Process improvement and operating models

I have been stimulated by an article by Craig Tickel of GoLean6Sigma that argues that there are only 5 ways to improve a process:

  1. Reduce Non-Value-Added Steps
  2. Improve the Measurement System
  3. Reduce Common Cause Variation
  4. Reduce Special Cause Variation
  5. Move the Mean to Improve Process Capability

I love people who simplify and Craig has some great exhibits in his article, so what follows is not meant to be a criticism of Craig, it is rather pointing out that a strategic/operating model approach to process design is different from, and probably should come before, a Lean6Sigma approach.

As I thought about Craig’s list, I came up with the following list of additional ways of improving a process that stem from taking an operating model view of the problem, as opposed to a waste and variation view.

  1. Digitize or automate steps in the process to reduce cost or improve value delivered
  2. Combine steps in the process with similar steps in other processes to get economies of scale
  3. Link steps in the process with similar steps in other processes to get good practice sharing or standardization
  4. Separate the process into two or more processes or sub-processes to enable better delivery of different value propositions
  5. Outsource steps in the process to suppliers with better skills or lower costs
  6. Change the suppliers or the contract with suppliers to get better supplier performance
  7. Change locations for some steps in the process to attract more skilled staff or to access lower cost labour or capital or other inputs to the process
  8. Change the people model or the people working on the process to get better people performance
  9. Change the decision authorities related to the process or to steps in the process to get better decisions made (e.g. centralize or decentralize or ..)
  10. Change the overhead activities supporting the process such as finance or HR or IT or … to reduce overhead costs or improve value delivered
  11. Innovate steps in the process to develop lower cost solutions or greater value or less variation (this last point is starting to overlap with Lean6Sigma work)

For those of you familiar with the Operating Model Canvas, I have been mentally working through POLIS (processes, organisation, location, information and suppliers).  I could go on (adding M for POLISM), but hopefully the point I am making is clear.  Lean6Sigma is for improving a process that has already been designed to fit with the strategy and operating model.  But often, what is really needed is a process redesign to get better alignment with strategy and operating model.  While Lean6Sigma goes some way to achieving this, an operating model approach followed by Lean6Sigma is more likely to get the job done right.

Posted in Design steps, Lean, Op Excellence and Op Model | 2 Comments

Lean6Sigma and operating models

I was stimulated by a LinkedIn discussion about how to ensure the customer is part of thinking about waste.  The answer of course is obvious – the customer is at the heart of all judgments about waste.  If an activity does not contribute to the value proposition that the customer is receiving, it is waste.  If the activity costs more than the benefit that the customer gets it is waste.

This then caused me to reflect on how to make some of these judgments especially in situations where the customer does not pay for the service being provided: a common situation between internal functions or in public sector activities and charities.  Here all of the tools of marketing and product development and service design need to be brought to bear: focus groups, indifference curves, observation, surveys, enthnography, etc.  Which is one of the reasons for the Lean guidance “Go to Gemba”.

I then began to consider what factors other than judgments about “customer value” and “cost” should influence judgments about waste – and where “strategy” comes into the equation.

Strategy defines who the target customer is and what needs of these customers are being addressed and what needs are not being addressed.  If Lean work is done without defining in advance who the customer (or beneficiary) is and which of these peoples’ (or organisations’) many needs are being addressed, it is easy for the work to lose direction.

But I also realized that there is another constraint on judgments about waste – the operating model.  Whether you consider the operating model to be part of strategy or not, it is still something that needs to be decided before work on lean process design begins.   The operating model is the collection of processes, types of people, organisation structures, IT systems, locations, suppliers and management systems chosen as the best means of delivering the value proposition.   When “leaning” each process, black belt experts need to consider how that process fits into the larger operating model and avoid eliminating activities that may not deliver to the customer of the process, but may be important to the integrity of the larger operating model.

In practice, this is achieved by recording the “stakeholders” of the process – not just the customer; and recognising that the process must deliver value to all of its stakeholders.  As a result, an activity in the process may be necessary even though it does not deliver value to the “customer”: it may be delivering value to another stakeholder.

A simple example is the annual report that companies produce.  This adds no value to customers – but is required by law, and adds value to shareholders and debt holders and those who make contracts with the company.

Posted in Lean, Op Excellence and Op Model, stakeholders | Tagged , , , , , , | Leave a comment

People Models

One of the problems I have been working on is a way of thinking about people models.  I have had two stimuli.  First my son worked for Mark Warner (the holiday company) one summer and I was amazed to learn that he was paid almost nothing (€50 per week) and that his minimum wage was made up with the cost of his accommodation and benefits, such as use of facilities.  Yet there were three people applying for every job and he loved his experience working with Mark Warner:  he also did a winter season with them.   This seemed to me to be an interesting people model. 

My second stimulus was that operating model work is about making choices.  So I began to lay out the choices that are made for different people models.  Choices about the sort of people to be hire – skilled or unskilled, motivated by money or by meaning, etc.  Also choices about the offer to these people – full time or part time work, above average or below average pay, etc. 

This thinking evolved into a four column format for laying out a people model.  I have since thought that maybe it should be a two column model.  But here is the four column version.

slide1

Lets start with the first column: the type of people that you want to hire into the skill group.  Each item involves making a choice.  “Contract” is about the type of contract that these people want.  “Location” is about where the people live.  The next exhibit shows the choices made by McDonald’s for staff in restaurants and Eden McCallum for consultants. 

slide2

There are different ways of laying out these choices.  The above is a crude Powerpoint with a misspelling of McDonald’s!   But the next visual shows a more attractive display of a skill group with three roles in it. 

slide3

A similar analysis and series of choices can be made for the “value proposition” to the people you have decided you would like to attract. 

slide4

And for the column “career path”

slide5

The last column on “culture” is more difficult to boil down into a few choice areas.   This of course is also true of the first three columns.  There are many choices that are not represented.  But the cultural values column is particularly difficult to illustrate in terms of simple choices.  So instead, I attach my summary of the people model for Mark Warner (my views not an official Mark Warner document!).

slide6

In developing this tool, I also realised that it makes little sense to create a people model for a company.  People models need to be created for important skill groups within a company.  yes a company may have an overall employer brand.  But it will not have only one people model.

 

Posted in People model | 3 Comments