all right well welcome everyone to today's process optimization Academy session where we'll be doing a bit of a deeper dive into the histogram capability that helps us identify and isolate process outliers and Improvement opportunities the histogram was something that we added in the Tokyo release and then enhanced quite a bit in Utah and I wanted to make sure that we got it highlighted because it's a very useful capability just a reminder that we've moved these Academy sessions to once a month in 2023 so I've sent a few cancellations but it seems there's still a little bit of confusion or a few folks who have been showing up hoping for an Academy I don't blame you these are kind of awesome only to find that there is no Academy that day and I apologize for the confusion there I just so everyone knows the updated list of Academy dates can always be found on the academy page on the community site all right let's get the logistics out of the way for our newcomers and those that have stumbled upon us uh one of our recordings for the first time so the purpose of these sessions are to help us all get better educated on servicenows in platform process mining capabilities we'll provide how to's practical guidance and use cases product updates and as a group just exchange ideas and collaborate so we all continue to get better how these sessions typically work is we have a scheduled topic that will usually cover for the first 30 minutes though I think today is going to be a little bit shorter it's a very concise topic then we'll open it up for questions on the chosen topic for that session we can open up for questions above and beyond the day's topic like more of a sort of an office hours type session these sessions are all recorded and eventually posted on the community and then finally feel free to post your questions in the Q a section or the chat I'll do my best to answer them here today and for situations wherein we don't have an answer for the question immediately we always try to follow up and post them to the process optimization Community or maybe use them as a topic for a future session this is the slide that says hey anything we say and do here today can't be held against us in the court of law it's also the slide that says hey as part of these conversations if we make any forward-looking statements about might be what might be coming in future releases you should take them as forward-looking statements and make no decisions based on them whatsoever because they're always subject to change and I'm assuming you've read the fine print uh for those that might have not met before my name is Dan Grady I'm part of the product management team here at servicenow for our process mining offerings uh what we'll do here today we'll just do a quick process mining 101 for the newcomers and the folks that are not really familiar with the the solution uh then we'll dive into how we should be using the process mining histogram so we'll run through what it is some common use cases and then we'll jump over to the instance and demo a few of those use cases we'll get your questions answered and then wrap up for hey if you haven't already started using process optimization here's how you get started and here's some additional resources for you as always the quick process mining 101 with everything we do there's a design and desired path in our minds for how it should work and how things should play out whether that be planning an event like these Academy sessions or a workflow and when you design things you design for both efficiency as well as completeness to provide the best experience possible for everyone unfortunately what we designed isn't always what ends up happening in reality the reality is not all work flowing through the process is going to take the optimal path and that's going to have a negative impact on the experience that both either employees or customers requesting service are having it's also going to have a negative impact on the people who are actually doing the work and fulfilling the requests unfortunately understanding the reality of what's actually happening within our workflows and our business processes and improving them isn't always the easiest thing to do what in platform process mining allows us to do is use the audit log data that is generated as work moves through a given workflow in the servicenow platform and we use that to create a visual representation of what is actually happening within the process and this new level of visibility is going to help us accelerate our ability to identify inefficiencies non-conformant process activities and Improvement opportunities process mining just gives us the ability to answer process questions that have historically been very challenging to answer so where traditional analytics will allow us to answer a lot of the what questions about our process process mining helps us answer the why so things like where is our process getting stuck where is unnecessary rework happening where are the tickets ping pong between groups where aren't we conforming to the process we've designed and these are all things that have historically been difficult to answer and act upon and now they're available to us in a couple of clicks on the platform with servicenows and platform approach we like to say that process optimization is going to get you to the why behind all of your kpis and in doing so that empowers everyone involved in the process to make more data-driven decisions about which improvements to make right there's no shortage of opportunities to improve your process having the data you need to actually prioritize which ones to act on first is really important all right on to today's topic using the process mining histogram so I'll run through a few slides and then we'll look at a couple of quick live examples first I thought it would make sense to Define what a histogram means so it's a it's a graph used to represent the frequency of distribution of data points of air one variable histograms they often classify data into various bins or range groups I always find myself saying buckets I mean we count how many data points belong to each of these bins or buckets the example on the slide here is just showing how many students scored in a certain range on a test now how do we apply this to servicenows process modeling or put it in context of process mining from the servicenow perspective well in our world the data point that we are classificing is not a student right it's it's usually going to be some unit of work whether it's an incident an HR or customer service case a life cycle event a request a change a field service order or maybe you've built a custom application they have different types of work maybe a legal matter that you're moving through and then what we're going to be putting these units of work in is potentially two types of bins ranges or buckets one is a duration bin and another is a repetition bin the duration bins usually represent the ranges of time between let's say a state or assignment group transition but it could be applied to any arc on a map and the repetition bins typically represent the frequency or the number of times a piece of work moves into a state or assignment group or really any node on the map on the workbench you're going to have the ability to choose which Ben bin or bucket you want to analyze for a certain transition or Arc or node that's out there now having access to this type of distribution analysis in the analyst workbench to help us pinpoint specific sets of work that are either long runners or bouncing in and out of specific States or assignment groups that's really helpful right whenever we see these types of things they're typically going to be Improvement opportunities or processing efficiencies the histogram is also helpful in validating some of the average duration values you see in the workbench right so you might see on the map you see the between the different nodes you have those arcs and you see some average durations well you know what where did the majority of the records fall in that that average range there are there a handful of long Runners that are potentially skewing the average um so it's it's good for validation as well as helping us start to isolate uh information so like I said this is ratio is really useful but in the Utah release we give you the ability to go one step further and start to isolate the map based on the data in the histogram bins basically we give you the ability to start drilling down to focus in on just those outliers that you want to focus in and again we'll look at this in action in the demonstration so the histogram is available by default and and can be applied to any and all use cases but there there are four primary scenarios that we see uh people using it for uh the first is looking at repetitions at repetitions at the activity node level so in the first example right you see on the screen here the slide here what we're going to be doing is we're going to isolate only the incidents that went into the awaiting caller info state three or more times right right so that's that's kind of like hey I see this state on the map but I really only care about the situations where that's happening more and once or maybe instead of State we choose assignment group as our activity and and we use this to focus on work that is touching a specific team more than once during the life cycle and maybe our process is designed say something like um hey once something leaves level one it shouldn't come back and if it does um I want to know about it and those might be good training opportunities or an opportunity to update a KB with more recent information or something like that now the second example is highlighting duration bins based on state transition time so in that example the average duration for work moving from work in progress to close complete was four days but you'll notice that if we look at the histogram we have 61 records that's somewhere between three weeks and five days in one year so so maybe we want to dig into those those specific outliers like what's going on with those 61 cases that one there might be skewing our average numbers there but also their opportunities to improve in the third example we're isolating records based on the duration of time between group handoffs so let's look at work that took longer than three days to move from group a to Group B and how many of those records potentially breach some form of SLA I mean then the last example is showing us how we can use the repetition bucket or bin on assignment group activity knows to isolate ping pong situations potentially again the histogram is an always-on part of the analyst workbench so your your activity definitions are going to drive the use cases and what the histogram will show but these four are probably the most common scenarios that that I see happening okay last thing before we jump into the instance and take a look at a couple of those scenarios live um I've mentioned this a few times but wanted to make sure it was clear the histogram was introduced in Tokyo but in Tokyo it was only available on activity and node or I would say that arcs that no transitions on the map not the actual activity nodes themselves meaning you'd only see it if you clicked on the arc on the map we only offered duration bins in Tokyo and there was no ability to to drill down so in Utah you'll see the histogram when you click on both of the nodes on the map they're like activity nodes as well as the arcs and you'll have both duration and repetition bins available to you and you'll also be able to drill down based on the Selections in the histogram so it's a it's a big difference in terms of the level of interactivity and functionality between Tokyo and Utah um and the other thing is if you go to your instance and you aren't seeing the behavior you expect for the release level that you're on what you should check is just check to make sure that your process optimization plug-in is up to date for the release that you're on right you want to make sure that you're using the most current the current plugin for for the release that you're on because these are front-end things that that sometimes you need to update the plugin to get all right now let's jump over to the instance and look at some of those examples live so the first one that we'll focus on was we'll just look at that example where we're we're looking at things that potentially are going into a node on the map in this case a waiting caller info uh more than one time so a repetition type example so I could come in here to my map and I will just zoom in and I'll find my let's say a waiting caller info node so these are my tickets that are going into wedding color info we can isolate that then I'll get my pop-up with the additional information over here and you'll see that I've got my histogram on what this is showing me that hey of those 8 400 incidents that went into a waiting caller info 7 200 of them only went into waiting caller info once and then 894 went into them that state twice but what I really maybe want to focus in on is anything in which it went into a waiting caller info more than three times that's a big problem area so let me focus in and I'll click on three four five six seven eight nine I have two records that went into awaiting calling from nine times that's crazy and I can come in here and I can click apply filter and that's going to update my map to show me only the records that went into a wedding coloring for more than three times so I had 200 or 349 of those and I can see here now that if I go back to the histogram it's only going to show me the data for those those three now I can start doing anything else a number of other things everything else that you're familiar with in terms of slicing and dicing this data Maybe I want to look at the channel here and I can see that I've got a bunch of these that came in Via self-service and portal definitely opportunities but phone like 39 of these came in via phone and I had to go back to the user more than two times to get additional information there's got to be opportunity there to improve that initial phone conversation potentially so let's just narrow it down to those I mean now I can come in here and if I wanted to I can see which groups or teams uh it looks like it support Americas is another area where the majority of those things are coming in so maybe I want to focus in on that team and how they're performing again once we use the histogram to narrow it down we can start using the rest of the process optimization workbench to continue to hone the information and look for follow-up opportunities so that's one looking at repetitions going into a node I'm going to jump over here to this HR case example to look at the next peace see a questions come in oh all right uh there was a question will these sessions be recording absolutely all the sessions get recorded and they get posted on the community site for follow-up for those so you can share it with your peers now next example uh we'll just look at this HR case example and in this again what we're going to do is we're going to look at maybe seeing uh long Runners that we want to focus in on using the duration bin so I'm going to come over here to my HR cases and I'm specifically going to focus in on these HR cases that go from work in progress to closed and I can see on average it's taking us about four days to move from work in progress to closed if I want to dive deeper into that and see like hey what subset of the data is falling in you know where's the majority of the cases falling in terms of time to close or time to move from work in progress to close and I can see that hey of those a good portion of these are taking less than two days to move through the steps through that step but I want to come over here and I see I've got 61 that took longer than three weeks and five days uh to move from work in progress to close so maybe maybe I want to focus on just those 61 long Runners that potentially are skewing that average a little bit um and narrow it down to just those and then I can start using again my breakdowns to look at these things and I see I got 12 of those that were referrals like what's going on with referrals so maybe I want to focus in on referrals and I can start to see all right ready to work in progress on average it's taking us one month to do the referral work what's going on there and then there is some reopening going on as well so there's some extra steps in the process so just an example of using the histogram to identify a set of long-running cases and then being able to narrow it down based on the long Runners that might be outliers so definitely outliers and skewing our results next example is I'll just come over here and I will use some do some group analysis group transfers that's a very popular use cases people analyzing kind of who's holding on to the tickets the longest or how work is moving through different teams inside of the organization I mean in this example here I'm going to just start with my finding a definition right we've had a whole session on these Improvement opportunities that we provide out of the box so when you mine your data it immediately calls out an opportunity to improve so I want to focus in on these these multi-hop issues so incidents that are going through two or more reassignments and I'll just narrow down the initial results that I'm looking at on the map and then if I want to expand this out a little bit just so we can see the data a little bit more I do have this one high volume handoff here these incidents or that are going from this it Support Service Group to it support Americans and if I wanted to focus in on this right what I can see now is I can see hey I've got both my duration and repetitions buckets on this one so if I wanted to focus in on just the tickets that took this trip two three and four times I could do that or I could start using my duration bucket to see hey maybe I only want to focus in on the tickets that took a long time to move between these two groups or to get handed off so if I come in here and I want to start looking at let's say anything that was more than three days took lower than three days to go from the first group to the second group I can do that here so just isolate those apply this filter and it's going to narrow it down to just the tickets that took longer than three days to take that trip now the next thing that you might want to do again we can expand this out a little bit to see if there's any whoops other place other paths that were taken and you can see it's going to a couple of other places but another way that you might want to look at this and I don't have the best data to show this but as if I came over here to my variation analysis and I said you know what maybe I want to sort this by the most records and I can see that the majority of these tickets that took longer than uh what did I say three days to to move between that first group and the second group they stopped right that was it you can see that 123 of them lasted they found their home in I.T support Americas and that's where they were resolved or closed um but you know what maybe I can start looking at my variation analysis to see if there's any examples in which these tickets went through that trip but then they wound up being closed someplace else and again my data doesn't support it but if you had a large volume of tickets that came in took longer than three days to go from one group to the next and then they went to a final group to get closed that potentially is an opportunity to improve your your routing organization it's like hey can I just get it to that last place um more quickly or the ideal Place more quickly and save us those three days so this is just a way that you can potentially use the histogram to identify long-running handoffs in the organization and then combine that with variation analysis to say hey there might be an opportunity to do better and get things to the right place more quickly and save large volumes of time all right and now just the last example that I wanted to show here let's bring this back to something that's a little bit more manageable is using the the histogram to focus in on just repetitive uh groups things that go to the groups more than once or a certain group more than one so let's say um if I find my let's say level one desk right perhaps you have a process that dictates things that go back to level one uh or maybe your process says hey once it leaves level one it should never come back and if that's the case right I might be able to use the histogram and the repetition bins to say hey you know what I only want I want to identify and look at let's say these 183 records or these records that go back to level one more than once in the process right again your process May dictate that it goes back but this is just an example of if you are one of those organizations that says once it leaves level one it should never go back there and I want to focus in on those so now I can use this to isolate down things that go to level one someplace else and then eventually that they're they're coming back to level one um and we can look at the groups in which our parents went back to level one often on the map or use the variation analysis again to start digging into kind of the scenarios where that's happening more than once so just four examples of different ways that you potentially could use the histogram to identify opportunities or inefficiencies in your process um there's probably hundreds of other ways that you could use this capability depending on which activities you decide to set up but it's really powerful very useful to help you start narrowing down and isolating opportunities in your data all right just going to bounce over back to the slides and we'll we'll pause here and see if there's any any questions does anyone have any questions I'm going to just check the chat um and see if there's anything that's in here please if you have questions post them in the the Q a or the chat area and I'll try to address them seeing a question here about someone was trying to use the histogram to filter for a few outliers and received an error that there was the records were too low well the number of paths was too high I've not seen that one before but I'm if you could send me a note I'm interested to follow follow up on on kind of why that might be happening or take a look at that example with you I I mean I've applied it myself to the uh to one record so I I I don't see where I have not seen that myself yet so I'm interested to see how that's occurring or why that might be occurring so if you if you shoot me a note let's let's schedule some time to follow up on that one um so the another question that's come in do we have the histogram for all tables that extend from task uh basically the histogram is going to be available uh to you for any data that you've mined with process optimization um so it really it's really a matter of hey is auditing turned on for that table great do I have the entitlement to mine that data with process optimization great um and then if you have both of those things in place whatever activity you choose to mine you're going to have the histogram available to you inside of the process optimization workspace so it's really it's it's not a matter of whether it extends from task it's more a matter of do we have the audit data for that table or is ordered and turned on for that table I hope that helps any other questions that folks might have all right we'll just come back and check in a second uh just for those that have not checked these out already you've come here this might be your first time a first exposure to process optimization you might be saying to yourself This was great but are there other resources available to me uh that would help me learn up a little bit more about process optimization or in-platform process Mining and the answer to that is yes of course we've got an on-demand a training course on the now learning site called process optimization Essentials self-paced course free uh it's usually about 90 minutes that's 45 if you do it at two times speed very helpful to help you get the basics we have applied process mining here at servicenow and pretty extensively now at this point and we've written a white paper about how we've applied process mining how we approach setting up a process mining program and then detailed some of our initial findings for a handful of workflows like incident change Asset Management um HR case I think is also detailed in that white paper I was really good Forbes article written by our our SVP of analytics VJ code 2 that focuses on the transformative power of process mining on employee experience so that's something that's worth checking out um you've probably found this session via the process via the the community site but if you're not familiar there's a specific product Hub focused on process optimization or in platform process money on the community form we post all these recordings there we post a ton of other content there it's also a great place for you to post questions that you might have because we typically check them every day to see if there's any questions and try to do our best to turn around and answer in a very timely manner usually within in the first couple of hours of you posting if not the day um as many of you have probably learned is process mining is a new muscle for many of us myself included and our customers I mean how you use the analyst workbench um to identify Improvement opportunities is is a little bit of a an art in some ways so what we've started to do is we've started to document a different use cases or common use cases or scenarios in something we call the use case Series so things like using process mining to help us identify work that takes longer to assign than resolve or how do we use process mining to help us do some SLA breach analysis or how do we use process mining to help us do ping pong analysis we've started to record those little brief five-minute recordings and post them in something we call the use case series that you can find on the process optimization product Hub on the community form and then lastly if you wanted to find a session that said this was great I want to kind of walk through the steps of what plugins do I need to turn on how do I create my first project and what are some of the building blocks for getting started within platform process mining there's a really good live on now session available in the community that walks you through all of those things basically like from start you know activating plugins to mining your first project in a matter of minutes so that session is a super useful way for folks to to start getting started all right last chance for questions uh okay someone's asking how do I find that uh white paper that you just mentioned uh that white paper is going to be available on the process optimization Community FAQ so if you go and you search servicenow process optimization FAQ that will get you it should be the first thing that pops up in your in your browser search you click on that and the white paper there'll be a link to the white paper in there you could also search for now on now process uh mining uh it might come up there but the the best way that I know how to get you there is to go through the servicenow process optimization FAQ and that page is going to have a number of links on it and one of it will be the URL for the white paper any other questions before we call it a day awesome well I appreciate everyone joining here today uh this was understanding and using the histogram session today now the remember only once a month we have these Academy sessions now so the next one is going to be scheduled for August 11th and what we'll be doing there is we'll be diving into the new enhancements that are coming in the Vancouver release um and then following that I've had a number of customers ask me about hey is there anything special that I need to configure uh process mining for a custom application that we've built with app engine on the platform uh so it's not a big lift it's really just a matter of creating because we don't have a Content pack for your custom application it's a matter of creating a process configuration record but I realize that we have not spent a lot of time on process configuration records during these Academy sessions so I think that'll be a good topic for for that session in September as always if you have ideas or topics that you want to spend us to spend time on please post them to the community or shoot me a note um and and I'm happy to add them to the list of sessions always looking for for topics that would be interested interesting to our customers last question coming in ah good question question came in about when do you use sample Mining and when do you use full mining um so just so for those that are not familiar um when you create a project to mine which is essentially you're essentially almost configuring the same thing as your your configuring a report right you're putting in some you're picking a table putting in some filter criteria and then you're setting what are your activities so those things on the map we use looked at States and assignment groups today um so you just pick the activities you want to include in the map and then your breakdowns which you saw slicing and dicing by Channel or priority or category um there so those it's really straightforward to configure your process optimization project once that's configured you have two options full mine versus sample mine um the the first time you mine always recommend that you start with a sample mine and then what that's going to do is it's going to mine about 3 600 records um that initially and really what that allows you to do is just make sure that the activities and the breakdowns that you configured are in line with what you expected them to produce right you know there's sometimes you you pick a field as a breakdown and you think it's going to produce one output and then you look at it and it's like whoops you know what I wanted something different or occasionally what you'll see is that you you pick a reference field as a breakdown but that reference field doesn't have the display value set on so you really need to kind of go one level deeper and pick the label or the display value from that reference table to get it to show up right so just best practices to to start with a sample mine just to make sure that your activity definitions and your breakdowns are showing up as you as you expect them to show up and then once you think you got everything right then you go for your full mine I hope that was useful any other questions that folks might have I got a question about the sample mine is it a random sample I believe it is just going to be the it's sequential it's going to be the first 3600 records that meet that criteria when it goes to the query but that's something that I'll double check and I'll post the response on the community I know I've I asked that question before and the answer I believe the answer I got was sequential uh the first not random but I'll double check and I'll post to community any other questions all right well with that thank everyone for their time we'll see you next month and uh if you have questions in the meantime please please post in the community and we'll get them answered thank you everyone