Monthly Archives: September 2015

When the domain experts (who will use the software) understand the software logic and are involved in its development

A critical problem with software implementations is lack of “engagement” from the people who are going to use it.

The term “engagement” might be accurate (ie to say, they aren’t interested) but misses a great deal.

From our experience working with the Aberdeen oil and gas industry, it seems more accurate to say, many professionals would rather pull all of their teeth out than have a discussion about software. They hate software, they hate their company IT department, they hate software companies and their stupid promises, they hate software which tells them what they can and can’t do, or software which tries to get them to ‘comply’ with something someone else decided.

Often – or perhaps usually – they don’t understand what the software is telling them, or why it says what it says.

It makes them feel devalued as a human being, spending their days trying to work out how to get on with sofware rather than actually doing what they were trained to do and want to do.

Once they’ve figured out how to use a software package, even if it took them months, then they really really don’t want anyone to take it away from them and start trying to get them to do something else.

So we get a situation where, yes, users are not very engaged with software development.

how do we fix this?



Expert centric software – what we value

Here are some things we value when creating expert -centric software:

  • When the domain experts (who will use the software) understand the software logic and are involved in its development
  • Experts able to follow goals rather than following processes
  • Experts have the best possible software working landscape, with the best company data readily available and easy to work with
  • It is very easy to enter data and stories
  • Software tools which motivate people
  • Software tools which aggregate large amounts of corporate data
  • Tools which distinguish fact, analytics, previous experiences
  • Tools which help you find out about previous experiences which were similar to this one and what happened

Expert job categories – who would benefit from expert centric software

Here are some expert job categories who might benefit from expert centric software:




Building surveyor
Construction manager
Estate agent
Facilities manager
Geographical information systems manager
Geomatics/land surveyor
Housing adviser
Site manager
Building control officer/surveyor
Building services engineer
Quantity surveyor



Aeronautical engineer
Civil engineer
Commissioning engineer
Communications engineer
Control and instrumentation engineer
Electrical engineer
Engineering geologist
Environmental manager
Exploration geologist
Health and safety adviser
Land-based engineer
Maintenance engineer
Manufacturing engineer
Materials engineer
Materials specialist
Mechanical engineer
Mining engineers
Petroleum engineer
Seismic interpreter
Site engineer
Software engineer
Structural engineer
Technical author
Technical sales engineer
Water engineer
Chemical (process) engineer



Commodity broker
Corporate banker
Corporate treasurer
Credit analyst
Financial manager
Insurance broker
Insurance claims inspector
Insurance risk surveyor
Insurance underwriter
Investment analyst
Investment banker – corporate finance
Investment banker – operations
Investment banker
Investment fund manager
Pension scheme manager
Loss adjuster (chartered)



Advocate (Scotland)
Armed forces officer
Arts administrator
Clinical research associate
Customs and excise officer
Customs officer
Doctor (general practitioner, GP)
Doctor (hospital)
Education administrator
Energy conservation officer
Government lawyer
Government research officer
Health and safety inspector
Health service manager
Higher education administrator
Lecturer (adult education)
Lecturer (further education)
Lecturer (higher education)
Legal executive
Medical physicist
Museum education officer
Nutritional therapist
Records manager
Recycling officer
Social researcher
Sports development officer
Tax inspector
Teacher (secondary)
Tourism officer
Town and country planner
Trade union research officer
Trading standards officer
Urban general practice surveyor
Waste disposal officer
Water conservation officer
Careers adviser (higher education)
Civil Service administrator
Licensed conveyancer
Local government administrator
Patent examiner
Regulatory affairs officer



Information scientist
Information systems manager
Information technology/software trainers
IT consultant
IT technical support officer
Systems analyst
Systems designer/builder
Systems developer
Systems programmer



Media planner
Agricultural manager
Catering manager
Charities administrator
Logistics/distribution manager
Management consultant
Personnel officer
Planning technician
Project manager
Quality assurance manager
Recruitment consultant
Fisheries officer
Fitness centre manager
Leisure centre manager


Market research executive
Marketing account manager
Marketing executive
Restaurant manager
Retail buyer
Retail manager
Sales executive



Animal nutritionist
Biomedical scientist
Field trials officer
Food scientist
Food technologist
Forensic scientist
Soil scientist
Advertising account planner
Biomedical engineer
Operational researcher
Product development scientist
Programme researcher
Research scientist



Courier/tour guide
Exhibition organiser
Hotel manager
Tour/holiday representative
Transportation planner
Travel agent


The Software for Domain Experts Manifesto

  • Creating business value
    • If experts have the right information presented in the right way, it creates enormous business value.
    • They can make better decisions, keep organisations healthy, and be more productive, and enjoy their work much more.
    • They can even generate entirely new business models.
    • Giving experts better data is surely one of the best answers to some of the world’s biggest challenges – (reducing the cost of oil and gas production, solving carbon, running government departments well, everything else).
    • If a problem can’t be fixed by giving experts access to the right data, it probably can’t be fixed at all.
    • Many experts in critical roles have a tiny percentage of the data which would help them do their jobs better readily available.
  • What experts need
    • Experts need the right data presented in a clear and consistent way so they can make the right decisions.
    • The data needs to show them where they are, what is happening, things they should be aware of, times when they were in a similar situation before and what happened.
    • Entering data must be as simple as possible
    • They must have a rough understanding of how the system works
    • The system must also motivate the experts – give them suggestions which hold their interest
    • People working together must also have access to the same data, so they can discuss what it means
    • Experts may be part of a process, but it probably makes sense to separate the process management with the challenge of providing them with the right data, otherwise it gets too complicated.
  • In order to provide this
    • They probably need ongoing service and continuous customisation of the software to changing needs. This is not something big companies are usually very good at.
    • The user interface will probably be web hosted
    • Building the right user interface, and delivering the right data to it, will also need domain expertise
    • There will be different interfaces for people in different roles. People in a chain will see the data of people before and after them in a chain; people in management roles will see the data of their subordinates.
  • Under the hood
    • Getting the right data to the right place is no trival matter
    • It requires a lot of very complex data translation work, which will all need to be paid for, from the value generated for the domain experts – which means the domain experts have to be supportive of the idea
    • Data standards will be very helpful, particularly in working out the intersections between streaming and fixed data
    • Data analytics are important – but the effort is wasted unless the right data can be delivered to the right place to someone who understands what it is
    • The software will probably be cloud hosted
    • Low code tools will help put the software together faster – and make it clearer to the developers how it works – and make it easier to update.
  • The commercial model
    • Expert users must be involved in development as much as possible – since they (or their departments) are paying for the development
    • Need to be space for independent companies of all sizes in every layer – the user interface layer, getting data to the user interface, underlying databases, analytics, cloud hosting, low code tools
  • Common errors made in creating Software for Domain Experts
    • Too much emphasis on ‘business process management’ and not enough on getting the right data to the right people
    • IT department taking a reactive mindset – fixing problems rather than creating the best possible system
    • Not enough willingness to do the really difficult work!

Why should mathematicians make software?

The programming world is full of mathematicians, logical people, engineers. But they might not be the best people to make software.

Making software which really suits an expert work is more like architecture, designing buildings people really want to live in – who have a detailed internal model of how people behave in buildings.

Or consider the people who create high street retail businesses, who need to create a sense of welcome for customers, and also maintain security – quite analogous to what software ought to do.

Interior designers can create rooms which people want to be in.

Ergonomic experts can create tools which fit the way people’s bodies work so they feel intuitive to use.

Writers and editors can deliver information to an expert the way they want it.

Advertising people can make someone excited about an opportunity.

The police know how to make sure people are aware of a problem.

These are all people who know, in their own way, how to develop a good environment for expert work, and there are plenty more designers.

What if software was designed in depth by these people – and then once it had been comprehensively designed, the mathematicians, engineers and logical people step in to build it?

What is the expert landscape?

Experts in all organisations need to be good at spotting problems, spotting opportunities, and figuring out solutions to the problems and ways to go after the opportunities. They also need to be good at pursuing goals set by others.

What happens if we try to imagine software developed from the start to meet that objective.

The software needs to have a very sophisticated ‘model’ of the expert’s landscape. By sophisticated we don’t mean technically complex, we mean accurate – in the way a model of a building can be accurate.

I don’t know what the most important information someone in an organisation I don’t understand needs when they sit down at their desk in the morning. A school headmistress might be most interested in exam scores, absentee levels, temperature in the building, the mood of the parents, the status of a specific problem, a visit from an inspector, a problem with the building, a staffing issue, I don’t know. What I’m saying is, most of the above factors could be made available digitally if they aren’t already, and someone could build software to deliver it.

So she sits at a computer, the computer says “here’s the information you are most interested in right now”, and “here’s something happening you might want to know about”.

Continuing the idea – every small company has business development and sales staff. What information do they want most at the beginning of the day? E-mail can deliver them reactive stuff (complaints from customers for example) but their job is fundamentally to grow the business not reply to e-mail. Perhaps there could be some relevant news about something happening in their town or their industry, or something a competitor is doing, or an opportunity to meet a key client in a conference. There could be relevant internal news about a manufacturing problem or advance. I don’t know.

But with a detailed insight into this expert’s landscape, it would be possible to build sophisticated software to provide the expert landscape.

Note that we’re not talking about analytics, BI or decision making tools here. This can be part of an expert’s data landscape, yes. But the core part of it might not be particularly sophisticated – go back to the headmistress example – just information about the temperature in the school hall.


Software for stimulation

Here’s an idea (well I’m not the first to think of it, but bear with me)

Shouldn’t enterprise software stimulate its users?

Let’s say, the role of an expert is usually to spot and understand problems, spot and understand opportunities, and develop solutions.

Historically software has been very good at (let’s say) anti stimulating people – the software version of the smoke alarm which goes off by itself – it gives you a list of things it doesn’t like (usually something wrong with the data), which basically gives expert workers some extremely tedious work.

What software could actually do is almost ‘tease’ the user by showing them some information it has found which look like it might lead somewhere – so the user is inclined to follow it.

Analytics tools could be analysing data all the time for something which the human might want to pay more attention to.

An interesting idea, shared by Paul Cleverley at a Finding Petroleum conference earlier this year,  is that search engines should aim to surprise the user, rather than bring them exactly what they expect.

Case management tools could act like a mentor to the user by being able to say ‘last time the situation was like this, here’s what the people did about it, and here’s what happened’.

This all requires sophisticated software but is easily possible with today’s technology. There’s just a lot of manual work involved understanding the expert’s landscape and the information available, to work out what would be helpful.

Analysing the markets of SAS

SAS, a US analytics company, has a customer base which could be a close match to where the ‘Software for Domain Experts’ concept would work. Let’s go through it. This is all taken from the SAS website.

Capital markets – risk management, fraud, capital management

Casinos – marketing analytics, pricing, gaming analytics

Communications – audience analytics, network analytics

Consumer goods – supply and demand planning, marketing + sales analytics

Defence and security – operational analytics, logistics, predictive maintenance

Health care providers – outcomes, care delivery

Health insurance – managing costs, customer analytics, risk

High- tech manufacturing – product quality, marketing, supply and demand planning

Higher education – strategic planning, data visualisation

Hotels – marketing, customer loyalty, pricing and revenue, operations analytics

Insurance – fraud detection, profitability, compliance

K12 (up to age 18) education – effectiveness

Life sciences – research and marketing

Manufacturing – Supply and demand planning, product quality

Media – audience analytics, ad revenue management

Oil and gas – analysing reservoirs, assets, risk

Public sector – effectiveness, fraud, criminal justice

Retail – marketing, customer insight, demand forecasting, pricing

Small mid size business – forecasting, business intelligence

Sports – fan engagement, operational analytics, player analytics

Travel and transport – revenue optimisation, assets, demand forecasting

Utilities – grid analytics, customer analytics, energy forecasting, revenue

It isn’t just the software – its the IT managers

Most people dislike being told what to do – particularly by software.

How much of the problem is in the software itself – and how much of the problem is in the people who configure business software?

We’ve all heard people talk about how you want to get data users to enter data in drop down lists rather than free text so it is easier to process later,

And there are various other tricks that software designers use to get the results they want from users.

Business users are not like home / personal users – they don’t usually get much choice in their use of the software, and they are being paid to do it. So how much incentive is there to make software for business users which is really fun to use?

For the entire business there’s plenty of incentive, if it means that staff are motivated, using much more of their brains and excited by their work.

But for the IT department, are they just another problem to solve?


Where does this concept provide the most value?

Where does the Software for Domain Experts concept provide the most value?

I think we are talking about domains where:

  • there’s an enormous amount of data, in lots of different sources, which people need to somehow make sense of
  • there’s high value decision making involved
  • the expert roles cannot easily be shoehorned into a process, and efforts to make them into a process have (at times) just made it worse.

We are not looking at domains where there isn’t much data, or the data is very structured already or can be put into a process.

In some fields, the data can be managed by professional researchers and academics writing up reports – and that can be the best way to do it. But we are talking about software and automated tools here.

So domains I think our concept works best in could be:

  • Very senior business / organisational decision making at any level
  • Any marketing / business development role – figuring out where to take your company next and working out if it is working or not
  • Government intelligence role
  • Oil and gas role, managing complex infrastructure or reservoirs with enormous amounts of data
  • Government policy making role – including running cities and transport systems
  • Running health systems

Possibly it could be used in purchasing and hiring decisions – where there is a lot of data available.

The amount of data available in every field is going up astronomically driven by sensors and mobile devices, so this concept will be applicable for more and more fields in the future.

There are probably plenty more fields it could add value in, I am maybe showing a bias to fields I have a better understanding of.