 All right, so my name's John Spangler. I'm here to talk about value stream mapping and analysis at under four bits a second, hopefully. The slides are available through my blog unifyingIT.net and probably through the conference here as well. I've been a technologist now for 22 years, starting off in manufacturing. That's my interest in background in Lean Six Sigma. I'm currently an ITSM process architect at DeVita. If you're interested, you can follow me via my Twitter account or my blog that I mentioned earlier. At DevOps conferences, we hear a lot about value streams, but what exactly is it and how can we use it to improve our delivery pipeline? So value stream mapping is a lean technique for optimizing your value stream, your end-to-end process flow. Here's a picture of a complete value stream map. We'll go through each component through this presentation. Let's start with vocabulary. End-to-end means from the customer's viewpoint, so we optimize the entire flow. In Agile delivery, it could be product backlog to deployed or it could be order to cash, for example. Next are process activities which are like high-level flowchart items. Cues are the items waiting for work to be done in them are regarded as non-value add time. Work in progress, I'll catch up my slides here or just let my slides catch up with me. So work in progress are items in the queue waiting for things to be worked on. While processing time is the value add time or something that the customer would actually pay us for doing. And then lead time is the total process time plus the non-value add time, like waiting in queues. So too often we focus on what we are doing and how to improve that, but the biggest opportunity is really what aren't we working on? What valuable things are sitting in a queue waiting to be worked upon? So reducing queues is by far our biggest opportunity to increase our velocity. So where do we traditionally see these improvements? For large enterprises, it's reducing whips and queues. For Agile development teams, it's establishing lean flow with operations. For DevOps groups, it's decreasing your cycle time. So where do we traditionally see these improvement? Oh, I'm sorry. So it's about focusing on the biggest process bottleneck. Theory of constraints tells us that just one area constraints our overall flow here just like the funnels here on this diagram. DevOps focus on continuous integration and deployment and that word continuous really means lean flow through your delivery pipeline. Let's look at our map and our process activities here. They're typically three to eight items in a value map. These activities are often determined by your current data sources and that's okay, we can always improve our data collection later if needed. Queues are represented by the yellow triangles with the number in the center, the number of items that are in that queue. Value stream mapping is a point in time tool. So either pick a date for your data or grab an average or median over a period of time. So for the same time period, add your non-value add and your value add data. Often value add time is not initially captured by your organization. So if this is the case, consult a team, perform the activity for an estimate. The 100 items in 90 days on the left hand side show good agile practice. This is exactly where we want our large queues because we maximize agility. If we clog up the pipeline with WIP we just end up slowing everything down and it makes it hard to expedite items. The monthly deployment window on the far right is the bottleneck here with the large WIP and queue as such improving any activity before that is just going to stack up work at that bottleneck. If we want to improve our delivery speed we really need to have to address that deployment bottleneck first. The efficiency metrics here show value add versus non-value add time. The 16% indicates we have a big gap with our customer demand rate. The 63% shows that we have a lot of opportunities to improve our work in process flow. So to recap it's always the customer's view of end to end. You should identify and improve flow at the bottleneck, never flood or starve the bottleneck and tolerate any flow and efficiencies outside the bottleneck. Smaller batches are generally better if they don't starve the bottleneck. And for DevOps teams don't forget to look at your retirement cycle times because that's going to affect your next iteration. Finally, use a tool called a cumulative flow diagram to monitor your flow over time. So five minutes really isn't enough time to learn value stream management but this is a good overview. I highly recommend the book Value Stream Mapping by Karen Martin and then there's also another book called Learning to See. It's written more in the manufacturing context but it's a little bit easier to digest. So no matter if you're looking to increase your deploys per quarter, month or day, your check list is here to improve your velocity. Like a race car it doesn't matter how much horsepower you have when your foot's on the brake. It's got to be on that throttle to actually deliver value. So finally remember that Value Stream Mapping is just a tool, it's not the solution. These resources I have listed here might be helpful with addressing some of those challenges that you're facing. You can find me at JWS Pro Services on Twitter, unifyingit.net for my blog. Also now said DeVita is hiring and lastly Deming Rules.