this is meant to be a introduction to pa performance analytics i want to get you all going in pa if you haven't used it before so off we go uh of course we have our safe harbor notice so i'm sure you've all seen this at some of the other knowledge events i'm not going to read it word for word but just know that it exists so here's our agenda for today so we're going to talk about data collectors indicator sources and breakdown sources how do they get populated how do we use them of course indicators and our breakdowns and then what do we do with this data so dashboards and widgets so continue on uh so a little bit about me uh so i'm a cloud uh tech architect delivery manager for accenture um i have spent uh the last nine years working on servicenow i actually started working on servicenow before they use city names so pre aspen they actually use seasons in years so spring 2011 is where i started uh i spent my first six years doing domain separation uh and if any of you are working on domain separation uh good luck i mean it's always ten times harder compared to like a single tenant but it's very challenging and fun i'm a five-year community mvp i've got itsm certs and you know several micro certs uh right now i'm part of a large team working on a global instance itsm hr integrations uh pa so i'm supposed to be on this account for six months uh we'll see um i've been focused on change management we didn't have anybody who could take over change management so i took over change management i'm helping out the integrations team and i'm doing service catalog work um we actually committed to doing 150 service catalogs so i decided to help out that team as well accenture if you don't know accenture so we're a servicenow partner one of the largest partners um we help out a lot of companies uh 95 in the fortune 500 and three-quarter and fortune 100 three-quarters of them are in 500 we have over five 500 000 employees i always refer to it as a small planet so uh but it's great company's been great i've spent most of my time internal i.t except for last october i switched to accenture and everything's been great so enough about me so pa what's the difference between pa and like standard reporting so reporting as most of you have already used always gives you a snapshot of the current status so if i run a report on say open incidents that's all i see is the number of open incidents i can't compare that to yesterday i can't compare that to last month so it always gives you a current snapshot a lot of companies build multiple reports where they show me all the reports for company a and then all the reports for all the incidents for company b um or by group by priority so you end up with all of these extra reports that sometimes you just don't need compared to pa pa is about enabling businesses to track their progress right against goals so let's identify issues in the company let's identify bottlenecks do we have enough resources do we have too many resources things like that pa i say requires a little extra planning but the overall goal will be to give you this beautiful dashboard that drives your business at a quick glance you can look at it and go oh yeah we need to fix this problem so where do we get started i put this slide in here because unlike our normal development cycle where we build in dev and maybe push to test or staging and then production all of our reporting data exists in production just like a normal report we need to build our pa reports in production so we do not build them in dev and migrate them always build in production if you have a brand new instance it came with the free version i put that in quotes because it seems to change all the time but it's the free version of pace loaded so you're welcome to utilize that out of the box and then you need to get with your stakeholders determine what needs to be measured what are we going to measure against right push back when they say everything because that's always a bad idea right um so you need to set specific goals on what you're going to measure so i put a couple examples like how long does it take to close an incident how long did it take to close an incident this month compared to last month we added more people so it should be faster is it faster is it slower things like that what percentage of incidents are resolved within my sla commitments so here's what we're going to talk about this is uh the architecture so we're going to talk about the data collectors we're talk about the source tables indicator source breakdown sources how they get populated from the data collectors we'll talk about an indicator and then the breakdowns that you can apply to these indicators and then after you have them built well what do you do with them so you create a widget and put them on the dashboard so this is the journey that we're going to take today so data collectors so data collectors are the engines that run out and collect the data they are responsible for populating the indicator sources and the breakdown sources we're going to use indicator sources and breakdown sources to build our indicators and thus our breakdown so if you have a brand new instance turn the data collectors on just before go live let them start collecting the data if you don't have a brand new instance and you just got pa it's no problem so you'll notice that there's two type of data collectors typically to type so daily which runs daily so we're going to capture all of the new incidents uh that were created all of the resolved incidents things like that and then there's the historic one so historics are available to run on demand so why on demand because if we've learned anything from college in high school history doesn't change so once we run the historic one once then we've already got all that data it never changes and we don't need to run it again now unless you've had your instance and again you just got pa no problem then you could run that historic one and capture all that data that's in the in the history um and even after you've set all these up i know i'm guilty of this i set them up they all work perfectly fine the jobs look great but over time you need to review the logs to make sure that there's no errors maybe you've added an extra breakdown and it's you're now bumping into the limits that pa can handle so you need to review the logs make sure there's no error everything's working good so here is a simple screenshot this is the historic data collector you can see in the middle of the screen it says operator is relative and then relative start is 60 and then days ago so this if we ran it would go back 60 days until the um one day ago and collect all the history so if it was a daily one we would say daily and it would go one relative start and one relative n and every day it would collect all the data and populate it again into the two tables that we're going to use for indicator sources and breakdowns so indicator sources so as i just mentioned these are populated from the data collector um make them very simple use clear naming i'm guilty of this myself i created data sources that were just named like what was what am i collecting here i totally forgot so i've listed a few that come out of the box like incidents close it's pretty clear what that's going to capture every day that's going to run and it's going to capture all of the incidents that were closed incidents new incidents open instance resolved very clear all incidents no no no no don't do things like that or incidents for service deaths we don't need data we don't need indicator sources and we don't need them specifically for a department um that's what breakdowns are for when we get to breakdowns so use the baseline sources if possible and minimize filtering we'll talk about filtering because filtering can be done in the indicator itself but minimize filtering conditions so here's an example of an indicator source you can see it's very basic it has a name again use a clear simple name it has a frequency this one is daily we have daily uh weekly bi-weekly monthly yearly fiscal year so we have all kinds of different frequencies you can use most common is daily right and then we have a fax table incidence and then open on today simple so this is going to capture all of the incidents that were open today when the data collectors run this will populate this indicator source table with all of the incidents opened today so indicators indicators are tied to a single indicator source not multiple so an indicator is tied to a single source and they can be generated directly from an indicator source that's an automated indicator you can apply extra filtering at this stage if you needed to so what do i mean by that so you've captured all the incidents that are new uh in a past life i worked for a company that was comprised of multiple companies even though we collected all this data on every company we only cared about one particular company so at this point i can add a filter that says company is x and whatever my company is and that will filter that data for this indicator what you want to try and avoid is doing extra filtering such as group is service desk or group to service desk and network and server i mean all that filtering can be handled by your breakdown which we'll talk about and then on your indicators you apply the breakdowns so your most common indicators are automated uh manual formula and indicator groups formulas are great so formulas like what if i wanted to get the average how many how many incidents did we close today versus how many were open so i could build a formula incidence uh resolved divided by with a little slash incidence opened and that will give me my decimal of what that number is i can then wrap it in parentheses and multiply it times 100 to get my percentage so again i can see very quickly are we losing ground are we gaining ground and then indicator groups are just a group of a number of indicators so here's an example of an indicator you can see it's got a direction minimized so if the number goes up is it better or if the number goes down is it better there's a checkbox for key you could mark this as a key indicator there's no logic behind it just that you can quickly identify your key indicators it's got a unit this one happens to be a number and it's got a precision zero could be two if we were talking about decimals maybe we want to go out two decimal points so it could be two it's tied to my in indicator source which is incidents.new and i have collect records checked and my aggregate says count and that a value one nil so what if we didn't get any records well then we're gonna set it to zero now i don't show you but uh i'll point my arrow at the additional conditions so the additional conditions would be where i could say company is x right just to limit the results that i'm getting back from my data collector from my indicator source and thus my data collectors at the bottom i have this one has three breakdowns priority category and assignment group we haven't talked about breakdowns yet but they are associated to this indicator and then i have two jobs there's a related list with a two so i have two jobs associated to this and again one is a daily and one is a historic so very simple very clean there's not a lot of fields that you need to worry about just to get started with pa so breakdown sources much like indicator sources breakdown sources are populated from the data collector the breakdown sources are basically collecting a set of records like for example incident state so we want to capture all of the incident states maybe we have to do some extra filtering such as incident states that are active and in english if we had multiple languages so our breakdown sources will generate our breakdowns so break down sources uh specify when specify which unique values to use so again filtering which incident states so a couple of examples i listed again incident priority right maybe just the active priorities uh if you happen to turn off uh you know p5 right incident state or incident age days so that's uh that comes around when we talk about bucket groups so here is a breakdown source so you can see it's got a name incident state that's a choice field we're capturing the sys id the table is incident the element is state the language is english you may not have that line if you don't have multiple languages and inactive is false or inactive is empty so here we're going to capture all of our incident states for our breakdown source not a breakdown yet but just in our breakdown source so very clean very simple same thing would be for priority if it was a group then the facts table would be the group it's this group table so breakdowns so breakdowns are a way to slice your indicator data right like i want to filter it by priorities i want to see i today we opened up 50 tickets i want to view how many were priority ones how many were priority twos you just apply that breakdown you click on the breakdown and it shows you the data instantly on your indicator so breakdown types most common automated manual breakdown relations which are child member of parents and then booker groups so booker groups you define the bucket groups like i want to see tickets uh incident tickets that are open between one and two days and three in five days and five and ten days and then greater than 11 days so you define the buckets that your tickets will show up in and the system will automatically put them there when you apply that bucket group don't apply unlimited breakdowns so keep your breakdown list short and don't apply unlimited breakdowns so here i've listed a couple of examples incident priority incident state incident age groups very simple you apply those to your indicators so here's an example of an automated breakdown so insert our breakdown source we talked about breakdown sources is incident state the key to this is at the bottom it says breakdown mappings so now the facts table is incident in our field is state so when we apply this breakdown to our incident records it knows what to look for incident and state according to the breakdown mapping so here is our indicator so this one is number of new incidents 38 so it says 38 new incidents uh for may compared to 58 that we had last month and you see in the left i've applied there is two breakdowns showing priority and categorization and i've clicked on category and now my breakdown is filtering it's showing me the records the 38 records by category instantly on the screen so maybe i have an issue you know i'm a manager and i look and i'm like oh why why are server calls spiking today or maybe i click on the priority and you know filter it by priority why are we getting so many p1 calls right uh you know there must be critical services down so quickly a manager can look at this and figure out what's going on in the business and react to it here's another view of an indicator this one's called the workbench and again this is all test data so here's one that says in may may 16th we had 72 new records you can see quickly across the top there's 39 in progress and four on hold and again as a manager i can quickly see the average open age right the average reassignment times are we getting the tickets to the right group to resolve quickly the average last updated time and the percentage of the tickets not updated in the last five days so you can quickly see what's going on with your business and this is just one dashboard you can jump to other dashboards to see other tasks that are going on so here's uh here's where we are in our architecture so we've talked about data collectors we've talked about our tables indicator sources breakdown sources we've talked about indicators and applying the breakdowns so now that we have all this data defined what do we do with it so before we continue i'll ask lisa if there's any questions that built up i haven't seen one yet come through on this chat but i will double check um we do have a couple uh from the community it's moses and it says 2am here in australia so hopefully they'll be able to come back and watch so yeah okay all right then we'll continue on here so so now we've got our data built so now we need to do something with it right so first we need a dashboard what's a dashboard michael you said minimizing filter conditions on indicator sources oh yeah when you would want to add those conditions yep so so minimize the filter conditions on the indicator is that the question yes so don't get confused by applying filter conditions to indicators compared to breakdowns so you can filter out records again the data collectors will collect all of the for example all the incident new records right but what if you only want to report on one company don't instead just say uh group is service desk um you know if you want to filter by one company no problem but don't try and apply breakdowns as filters use breakdowns i hope i explained that correctly you said uh but what about don't apply unlimited breakdowns yeah so unlimited breakdown so try and minimize the breakdowns you apply to an indicator so pa has limits on the records that it will return in the data collectors i think initially 50 000 but for every breakdown that you add it goes up exponentially so what happens is maybe you are you know oh i'm only querying 10 000 incident records for example but i have 20 breakdowns associated it goes up exponentially and before you know it you're bumping into the limits of pa and you get the errors so if you're not reviewing the logs you won't see the errors so always try to minimize your breakdowns if possible otherwise again i think by the time you apply all the uh breakdowns the limit is a million records it can return and it changes all the time all right all right so so dashboards so dashboards very similar to home pages except much cooler so dashboards are a collection of your widgets dashboards contain both indicators and breakdowns and filters dashboards can have multiple tabs which make them so much cooler than home pages home pages don't have multiple tabs right but the dashboard does so from this view i just filtered show me all the incident dashboards and you can see this is baseline data so you can see that it comes out of the box with seven different dashboards and i would bet each one of those has multiple tabs to show you what's going on in your business now from this view in the upper left hand corner there's a new button i could create a new dashboard right from here if i'm in a dashboard for example then i could just click on the little hamburger and say new dashboard if i wanted to i could delete that dashboard maybe i made a mistake or i could duplicate a dashboard to use in the future so it's very easy to create a dashboard either from the new button from the main dashboard screen or from inside a dashboard so now you've got your dashboard now we need to apply our widgets or widgets are nothing more than a reusable view of an indicator you know very simple widget types time series right obviously it's a vision over time a single score a list widget type a breakdown a workbench which i've already showed you what a workbench looks like or a pivot so to add a widget to a dashboard there's a little plus sign you click the plus sign you get a menu that looks very similar to if you're adding content to a home page you choose performance analytics and then you're presented with what do you want to add a breakdown a list a pivot a score etc and you click on what kind of indicator did i build so if you clicked on score for example you're presented with your score indicators and you choose the indicator that you want to add to the dashboard and again you position it just like you do on the home page nothing different and that's how you build your dashboard and present your information very simple you know and that's really what this talk was about is to get you going into pa it's not detailed in talk it's just to keep things very simple and keep it very organized so now we've talked about everything our data collectors our source tables our indicators are breakdowns we've got all that data now we built our widgets we put our widgets on our dashboards simple i hope you agree it's simple so lessons learned identify the metrics that matter for your business right your business objectives i mean and again push back when they say we want to see everything right it needs to be specific business objectives uh try to use the baseline indicator sources um you know you'll notice that as i as i showed you indicator sources like incident new incident open into resolve they all come baseline so unless you have a custom table you don't need to really create any indicator sources your indicators dashboards breakdown don't need to be perfect especially in the beginning right just get the data going the data collector will always collect the data but maybe you applied the wrong breakdown maybe you've sliced it the wrong way it's no problem the data still exists you just need to readjust your breakdown or readjust your indicator it doesn't have to be perfect day one and then keep it simple i mean i'm i'm guilty myself of not keeping it simple so i mean that's a definitely a lesson learned for me and then content packs so if you have the full pa you can install uh there's lots of plugins with content content packs and all of the necessary pieces the indicator sources uh the indicators all come in that content pack so initially there may not be anything for you to build it may all come in the plugin and so with that we do have a question so json posted do you have examples for using pa with the main separation i don't have any examples of using pa with domain separation um but i i think it's brand new in orlando if i'm not mistaken i think it supports it now but i don't have any examples of that when i worked in domain separation um pa wasn't even around unfortunately do you have any other questions coming up [Music] and okay he says i know it breaks out in the simple objective of this talk until it breaks out of the simple i didn't hear you i didn't i don't i don't know what you're you say it again um yeah it's a js wrote um in a reply i know it breaks out of the simple objective of this yes yeah i'm sorry i'm sorry that you're on domain separation but i mean again you know you can get really good at it over over time i mean i i think i just spent way too long in it well okay well you know our 30 minutes is up and a little over but this has been great if anyone has additional further questions um please do reference that this link uh that is available in the chat um the recording group is already there as soon as even when we stop it and you can post more questions and michael will be following it to be able to answer questions in the future so again we thank you all for joining us and have a fantastic day and knowledge thank you michael thank you take care everybody