- Zsolt Berend
About 85% of the matter in the Universe is unseen matter. We can only observe it by its gravitational influence.
What is the % of unseen work & queues (waiting) in your organisation, and what is the impact on flow efficiency?
Flow efficiency = (sum of time that work is being worked on) ÷ (total end to end lead time) in percentage. E.g.: If the total time that work is being worked on is 5 days and the end to end lead time is 50 days then flow efficiency is 5 ÷ 50 = 10%. This means that work is waiting 90% of the time, which is often the case at large service-based organisations! This is shocking and often organisations are not aware that work is sitting idle in queues for up to 90% of the time, before generating learning & value. This is where significant improvements can be made to optimising for sooner delivery of value to customers The long wait times show that there are impediments to flow which could be caused by e.g. role-based silos, multiple hand-offs, component team silos, departmental silos, or too much work in parallel.
Answering these questions could be a challenge. Usually there is no end-to-end visibility of the flow of value, and no knowledge of queues where work is waiting. Focusing on flow though cannot be at the expense of quality value, or colleagues’ happiness. More details on this in section #4.
So how to get there? In this blog I will share some patterns that I have seen working at many organisations.
1. Install end-to-end transparency
Make work and queues visible across the entire end to end flow of value, from concept to getting value into the hands of a customer.
Value stream mapping can help identify all steps, work (outline business case, prioritisation, quarterly planning, …, design, …, code, …, testing, …, deploy to production) and queues (waiting for sign-off, …, waiting for analysis, …, waiting for integration test, …, waiting for deployment), handoffs, and information and learning silos. The number of steps varies based on the context, but could easily be in the range of 20–50.
Representatives from each state of the workflow, from concept to production, should take part in the value stream mapping. The exercise could take two or more hours and will often require two or three long tables, a giant wall or virtual equivalent and a large number of sticky notes. It is always a great learning exercise, as participants inevitably realize how little they know about the end-to-end flow of work, who does what, and how long it all takes. The session installs transparency of silos, queues, impediments, and a number of improvement items to act on. It typically highlights bottlenecks, long wait times, and too much work in progress (WIP).
Once work and queues are identified, they need to be reflected on a board. I often see teams creating rather simple boards, with just To-do, In Progress, and Done states, and yet in reality they work in silos with multiple hand-offs. Such a simplistic board prevents the identification of impediments, end-to-end measurability of flow, and WIP.
2. Connect work end-to-end
Not only there are multiple hand-offs, information and knowledge silos but also it is often the case that work is captured and tracked in various non-connected tools, different instances of the same tool, or in the same tool but with no linkage end-to-end.
Many times I have seen teams declare an item as done, when it only means that the unit testing is done. The item of value then waits for a code drop from another team to merge, run integration testing, user acceptance testing and code deployment. However, these steps are not tracked as these are outside the scope of the team board.
Another example is when shaping, analysis and design work is captured and tracked in an isolated board with no connection to code, and test work, and the rest of the pipeline. When you talk to the analyst team and the development team, they will both tell you that they get work done in two-week intervals. However, when you ask how long it takes end-to-end, they have no idea. There is no linkage between analysis and development so no knowledge of the amount of time in waiting between the two states, never mind time waiting end-to-end.
I once worked in a bank treasury division with teams that run stress tests with steps like scenario analysis, risk modelling, computer simulations and policy response. All work was maintained as individual projects with individual project plans and executed by individual teams, with no end to end visibility.
Thus, it is essential that the design of any boards that capture the flow of work needs to be end to end. Logical linkage from left to right needs to be established. Once there is linkage the measurability is unlocked.
A good proxy measure is the % of measurability within the organisation, the % of work that can be visualized and measured end-to-end. In practice this can be the percentage of work items that are linked to a release in production in an IT context.
3. Be comfortable with unlocking measurability
This journey of unlocking measurability requires leaders (at all levels) to establish an environment of psychological safety. In this context it means that there is no fear of revealing and sharing data. Data is not used as a stick but for learning and continuous improvement. Transparency and safety are required to establish any measure for learning.
As early steps in the journey I often see push back from e.g. service management to use the production inventory of releases. They argue that they are the experts and teams should only use the on-demand reports otherwise they would misinterpret the data and will lead to confusion. However, this prevents self-serve discovery, democratisation of data, and end-to-end measurability of flow. The underlying cause is lack of psychological safety and the fear that it will be used for reporting, it might reveal inconsistency in the data structure and inaccuracy of data.
This presents a great opportunity to bring together representatives from all the potential silos across the end to end flow (e.g. biz, dev and ops) to talk about unlocking measurability and use it for learning and continuous improvements.
4. Create a self-serve dashboard
For continuous learning, there needs to be data and insights constantly available in an auto generated way. It’s the feedback loop which enables agility. It is essential to create a dashboard in order to make the data feedback loop transparent, to enable decision making and pivoting, in order to optimise for outcomes.
A dashboard should be self-serve and accessible across the organization by teams and leaders at all levels. The best way to achieve this is to create a data engine that automatically harvests data from all the data sources across the pipeline.
You can use that data layer to establish and visualize vector metrics, trends over time, and aggregate at different levels, such as application, team, product, value stream, and organization.
Treat the dashboard as a product: establish hypotheses, run experiments, and collect feedback on how useful the provided metrics and services are. Start with early adopters and innovators, and keep sharing their stories, social proof. It will attract the early and late majority. Start small, with one source system at a time, and keep expanding.
Establish regular show-and-tells, drop-ins; provide training and workshop services for teams and leadership to learn how to self-consume data, how to slice and dice; and draw insights. Provide coaching services for teams and leadership on measures for learning. Why is it important to measure, and what to measure?
Our recommendation is to measure Better Value, Sooner, Safer and Happier outcomes.
Better Value Sooner Safer Happier measures include:
Better: Better is quality. Depending on your context, quality measures could include system outages, time to recovery , error rates, reconciliation breaks, rework, and so on.
Value: Value is unique to your business and is measured via Objectives and Key Results, specifically the KRs, which are leading (customer behavioural pattern changes) and lagging (impact) value measures. This could include leading like subscription, awareness, acquisition, activation, conversion, retention, #site visits, #complaints and lagging like revenue, market share, profit margin and more.
Sooner: Sooner is time to market, concept to cash. It’s the lead time from starting work on an item of value to getting it into the hands of a customer. Another useful measure here is throughput, which is a count of the number of items of value over a given time period (which goes up as lead time comes down), and it’s flow efficiency, which is a measure of the percentage of time that work is waiting. This is one of the most important things to focus on. Look where the work isn’t. Alleviating impediments to flow enables lead time to reduce and throughput to rise.
Safer is Governance, Risk, and Compliance (GRC). It’s cyber, fraud, anti-money laundering, data privacy, not leaking your customer data, and avoiding news headlines. It’s agile, not fragile. In the context of software, measures include the percentage decrease in compliance breaches, such as mandatory risk stories not implemented, software binaries released without a link to the formal control environment, and safety incidents. It is also worth tracking the approximate percentage of time that Safety SMEs spend on reactive work (fire fighting) versus proactive work (building safety in up front).
Happier: Happier covers colleagues, customers, citizens, and climate. More engaged colleagues, more satisfied customers, and corporate social responsibility (CSR) to benefit both society and the planet we live on. Measures can include customer Net Promoter Score (NPS), employee NPS, CSR outcome, and climate measures such as being carbon negative and reuse as per the renewable economy.
Unlocking measurability is a journey that requires a pivot from fear of revealing data to sharing, democratising data; requires installing transparency of work and queues end-to-end; a shift in behavioural patterns to record, track and connect work. If there is a desire for agility, it is essential to have the data feedback loop established and available to all.
Although this journey is hard and non-trivial, it brings disproportionate benefits as it surfaces gold information and knowledge, leads to improved collaboration, continuous learning, and ultimately improved Better Value Sooner Safer Happier outcomes.
by Zsolt Berend, 11/10/22