Organizations cannot ignore today’s ongoing digital transformation. With every company now becoming a software company, if software isn’t being created correctly and quickly enough, industry pundits argue, companies are going to find it difficult to stay alive.
“There is a digital transformation going on and if you are not changing the way you are doing things in order to deliver software better and quicker, you are going to lose competitive advantage and become irrelevant as a company,” said Lance Knight, COO of ConnectALL, a value stream integration company.
While there have been all these methodologies applied to creating better software faster, like Agile and DevOps, organizations are quickly realizing these are not enough. “It is a technology wave, it is a digital disruption. All these things you need to be able to deliver software quicker, but just because you can program it and have it ready doesn’t mean you are getting it to production quicker,” said Knight.
What Agile and DevOps has enabled teams to do is start to think end to end, but there has been a piece missing to achieving clearer results, according to Mik Kersten, CEO of Tasktop, a software tool integration company. “We somehow ended up pigeonholing everything we measure and everything we focus on. It is siloed thinking. Agile was never meant to be a silo. DevOps was never meant to be a silo. They were meant to be around value streams,” he said.
Value stream is the flow of work throughout the delivery process. “If I don’t look at the system from one point to the other, it doesn’t matter how quickly I can deploy. I can deploy every minute, but if I am not getting something new every minute does that really matter?” asked Knight.
Value stream aligns DevOps and Agile transformations with the business in order to uncover areas of opportunities for improvement, according to Eric Robertson, VP of product marketing and management at CollabNet VersionOne, an Agile planning, DevOps and VSM software provider
“Agile and DevOps produced great technical outcomes that efficiently got out products and produced output, but the question is was that output or release meaningful to the business? Did it drive any meaning or help achieve the type of outcome the business needs in order to be successful?” Robertson said. “It is great to get to deliver products sooner, and it is great if you can deliver them safer given compliance and aspects. It is also great that the quality is higher and you have nice test tools, but in the end if [the software] is not meeting the customer need or the business objective it doesn’t matter. You can’t say you are delivering value. You are delivering output, you are delivering something. But you are not driving value,” he continued.
The value stream lays out how everyone in the business is delivering value and what their role is overall. “Value stream from our standpoint is really about the business side understanding and identifying business outcomes,” said Robertson.
“At the end of the day, companies are looking to move faster, but not lose any stability. I refer to that as speed and direction. Customers want speed with direction, and that is really what value stream management provides. Agile brought a lot to the equation in terms of optimizing the development cycle, but it was pretty narrow. DevOps expanded that point of view, but the focus for DevOps ended up being around automation,” added Brian Muskoff, director of DevOps strategy at HCL Technologies, IT and digital solutions provider
Additionally, there have been many shifts happening in terms of architecture approaches, such as microservices and the move to the cloud. In order to properly track features, epics and progress, teams need to understand if they are improving development and delivery velocity.
“We demand visibility and transparency from every other part of the organization; value stream management lets us put a methodology and system in place that gives us transparency, visibility and governance independent of the level of automation and independent of the tooling that is there to help us start improving work that is going on in these teams,” said Jeff Keyes, director of product marketing for software company Plutora.
Getting the value out of value stream management
In most organizations, software delivery value streams are grown organically. Throughout the years, organizations have documented and changed up their workflows. However, their organic value streams weren’t created or crafted with the notion of how things can flow easily, nor grown with efficiency in mind, according to ConnectALL’s Knight.
He explained that in order to understand the value stream, it starts with education. “If I were going to start today, I would really try to get as much education about value stream mapping, lean principles, waste and study systems thinking, so I can look at these things and do an optimization exercise, analyze the value stream, map it out and decide what tools I want to put in place based on that,” he said.
Once there is comprehensive understanding, HCL’s Muskoff explained the best practice for getting started is to start small. “It is pretty much a best IT practice in general. You want to start small, get some wings and establish a benchmark. Then, try to make some improvements day over day. That is a great place to start.”
Improvements should continue to be applied as an organization journeys down the value stream. “The team has to want to get better, and the only way to do that is to take an honest look at your practices and identify areas of improvement, focus on bottlenecks, and exploit those bottlenecks to increase flow,” Muskoff added.
CollabNet’s Roberston said beginning with objectives and key results (OKRs) can also help companies quickly identify outcome hypotheses and provide more insights into what they are trying to achieve down the line.“That input of taking, identifying and prioritizing those business opportunities will help you start planning, start creating, start delivering and also be able to pivot and learn,” he explained.
Additionally, businesses should be able to break down those outcomes and objectives and individually map them through the value stream. Value stream mapping enables users to understand what to look at, find bottlenecks, look into why things are taking so long, and remove waste, according to ConnectALL’s Knight.
According to Tasktop’s Kersten, in order to get the most value out of the value stream, you should start with the customer. While everyone is jumping on the notion of value stream and how important it is, the way it is being translated into customer results and thought about with customers within a large organization is “completely rife with confusion,” he explained. The problem is that organizations don’t know where a value stream starts. “Everyone knows where it ends — with running software — but the whole point of the ‘project to product’ transformation that we are seeing at the forefront of most organizations now is the value stream starts with the customer,” he said. “You need to be aligned with a customer goal. If your value stream doesn’t start with a customer, you are doing something wrong. You have to measure what the customer is seeing in terms of delivery.”
It is not only external customers that organizations need to look at, Kersten explained. There are many internal value streams and internal customers that need to be considered also. If you are not treating your internal services as their own value streams, “with its own roadmap, with its own resources, instead of teams behind it, you are not going to help those customer-facing apps. If you are not treating your value stream network, your toolchain itself, as a product, you are also not going to get the kind of results you are seeking,” he said.
And don’t boil the ocean with millions of metrics, HCL’s Muskoff warned. DevOps Research and Assessment (DORA) has provided four key metrics organizations can look at to help bring bottlenecks and waste to light: lead time, deployment frequency, mean time to repair, and change fail rate. Muskoff does note that each organization is going to have their own key performance indicators that are important to them, but the four DORA metrics are a good place to start. He also warned about treating all data equally. For instance, lead time isn’t a singular index across all teams, so be careful when drawing conclusions from the data.
The right tools can help provide that visibility, but it needs to be able to see what is happening across the entire portfolio, Plutora’s Keyes explained. “If you can’t see what is happening in one pipeline, then you can’t see what is happening across the portfolio. You need to have visibility into dependencies and related activities,” he said. Tools should also be able to take into account heterogeneous methodologies because teams will be doing a mixture of Agile, DevOps and even waterfall. It also shouldn’t matter what type of tool a team is using. Data should be standardized across the value stream so you can understand how to relate it all and see what’s going on. “How am I doing in my environment utilization and how relistent are my environments? Which environments should I move to the cloud? All these are great questions that I ought to be able to ask of a system and have it pull the data, use the data and go on,” Keyes said.
Value stream in 2020
As value stream moves into 2020, more tool vendors are going to take interest and figure out where they fit in the overall pipeline. “2020 will be the year where all these tools start to realize I have to be hooked into the overall toolchain, not from a toolchain perspective but from value stream management,” said Plutora’s Keyes.
2020 will also be the year organization’s start to see more results, according to Tasktop’s Kersten.
SD Times declared 2019 was the year of the value stream as organizations started to turn their focus on how to better drive value, and a majority of the year was spent on getting through the hype and finding clarity around what value streams really were and how they are going to have an impact, HCL’s Muskoff explained.
Now that there has been more research and understanding into the space, in 2020 “confusions will begin to get unwound, and some of these practices, definitions, and the way to apply value streams will become clear. It will go from this interest and start of the hype to actual company results,” said Kersten. “Organizations want to move from project management practices to become product oriented innovators, and they will start understanding how implementing and measuring product value streams are critical to them.”
These understandings will come from experimentation, and more experience with the value stream. According to ConnectALL’s Knight, most companies are not currently at a place where they are getting value stream right. There is going to be some trial and error before they get to a place where it is working, and companies will need to turn to consultants to help them improve. “What we are going to see is value stream managers start to come to fruition a little more in particular companies. We are going to watch companies look at how their value stream grew organically and they are going to try to improve that. We are going to see a bunch of niche businesses that are keen on value stream management come out and help companies achieve better flow and deliver software quicker,” said Knight.
Also expected from value stream management over the next year is the emergence of artificial intelligence and machine learning. According to CollabNet VersionOne’s Roberston, the value stream is increasing the amount of data businesses are receiving, and they are starting to get good at assessing the value they are delivering, but the next step is to really understand what that value is. For example, when trying to optimize the backlog and decide what to work on first, an intelligent layer can be added to the backlog to understand and assign business value levels or point to work that needs to be done and created. “Being able to utilize machine learning to help make work not only easier as far as the delivery, development and delivery aspect, but also ensuring that what is delivered is optimal for the business,” said Robertson.
HCLs’ Muskoff agreed that machine learning and artificial intelligence will be the next step in value stream. He predicts ML and AI are going to be able to go even deeper into solutions and uncover even more bottlenecks. “We are at the stage with the technology where dashboarding and KPIs are important, but to really get to where we need to go the tool needs to tell you where to focus,” he said. Areas where he believes machine learning and AI will be applied is bottleneck detection, planning, and predictive analytics around delivery time.
“More organizations will realize this isn’t optional. This isn’t just for technology visionaries. This is survival. If you don’t know your product value streams then you have no place in the age of software,” said Kersten.
The emerging role of value stream managers in software delivery
Managing the value stream has to be a human process, according to Lance Knight, chief operating officer at ConnectALL, a value stream integration company.
Part of a successful value stream is having an analysis in place where you map, measure, and look how things flow throughout delivery. “You can do a lot of stuff with a tool, but the tool isn’t going to do software or value stream analysis for you. It is a human you need to go in and look at it. There is no tool that is going to go out and look at all the things you do in your value stream and tell you want to do,” Knight explained.
Value stream manager is an emerging role in software delivery that aims to add the human element to value stream management. The value stream manager is in charge of making sure everything can flow, removing impediments, and getting releases out the door.
According to Knight, a value stream manager has more of a product owner or project manager background and will look at how work gets done, and try to make it more effective.
Brian Muskoff, director of DevOps strategy at HCL Technologies, an IT and digital solutions provider, said a value stream manager also needs to come from a technical background like a developer and tester as well as have general management skills. “They need to be a good communicator, organized, have attention to detail, have the ability to take a current situation and identify areas of improvement, and actually deliver on those improvements,” he said.
HCL has been experimenting with its own value stream manager internally, who they say is essential to their process. The HCL value stream manager takes on the role of Scrum master or release manager, being in charge of teams to get product shipped out the door, but also focuses on continuous improvement or process improvement. “The value stream manager is very much a necessity,” Muskoff said. “We are evolving. We are recognizing the importance of process improvement and formalizing it.”
Putting the value stream together
There are many different components that make up the value stream process. Businesses need to be able to map their value stream, analyze it and manage it.
Value stream mapping refers to the practice of looking at all activities throughout the delivery life cycle and mapping it out, according to Mik Kersten, CEO of Tasktop, a software tool integration company.
“The approach we have taken with value stream management is implementing and measuring those product value streams and doing so end to end, making that a core part of our management model, operating model and toolchain,” he said.
Value stream management refers to the process of managing the value stream. “Watching things flow, looking for areas to remove impediments and actually move things through and managing your value stream is a human process. Part of managing your value stream you want to do things like value stream analysis. Value stream analysis actually includes doing a value stream map, measuring it, and looking how things flow through your system of delivery. The exercise for all that I call value stream optimization,” said Lance Knight, COO at ConnectALL.
Jeff Keyes of Plutora, explained value stream management is not monitoring, it is not a feature management system, it is not a bug management system and it is not a build system.
“Value stream management sits as a way to interconnect the entire toolchain under one umbrella to basically create a framework or an ecosystem that the tools can plug into to make them all work together. It serves to interconnect disparate tools, creating scenarios from teams that don’t normally talk. For instance the help desk doesn’t normally talk to development directly, but using value stream management they can, should, and do,” said Keyes. “Helping bring alignment from the business to what is happening in the development world is where value stream management really brings this all together.”