Reporting vs Performance Analytics: What's the Difference? Part One

The reporting capabilities within ServiceNow and where exactly Performance Analytics (PA) fits in can often be considered confusing. It may be that you’re dipping your toes into this topic for the first time and therefore have no idea how to report on ServiceNow data. Alternatively, you might already be working with reports, and wondering if your organisation needs PA to achieve certain metrics that you’re struggling to produce.

Within our Reporting vs Performance Analytics series, we’ll be doing a deep dive into the comparison between the in-built ServiceNow reporting functionality and PA: how they weigh up against each other and which is potentially the best fit for your business. We’ll also go through some commonly asked questions surrounding this topic, and there’ll be helpful diagrams along the way. 

This is part one, where we’ll be analysing what’s possible to achieve within ServiceNow reporting, and how this compares to PA. 

How exactly do the two capabilities differ?

A commonly made mistake is that Performance Analytics is a premium version of the existing reporting capability within ServiceNow. The truth is, PA and reporting are separate capabilities that have their own distinct benefits and ultimately complement each other. 

Within the license subscriptions for ITSM, a light version of PA is included for free. This unlicensed version is limited to a set of pre-built indicators (think KPIs/metrics) and dashboards for Incident Management only, and restricted to 180 days of history for your trend analysis. The idea is that you can try out the functionality first in order to realise its benefits, and then decide if you should upgrade your license entitlement.  

The licensed version is called Performance Analytics Premium, and amongst other things, unlocks the capability to: 

• Create your own indicator calculations (KPIs/metrics) 

• Modify the OOTB (out of the box) content to apply your own additional filters or calculations 

• Create additional breakdowns (common data dimensions to slice and dice results by) 

• Extend the history of your results beyond 180 days

What can I do within ServiceNow reporting?

The ServiceNow platform includes a highly comprehensive range of visual charts to cater for different organisational needs. We won’t be going into detail for each one, but some examples include: 

• Single scores (great for dashboards) 

• Gauges and funnels 

• Bar charts, pies and donuts 

• Single and multi-level pivot tables 

• Lists (raw data) 

• Time series (trending over time)

Reports can be created centrally within your organisation, and shared to specific roles and groups. However, ServiceNow is predicated on a self-service model, meaning the reports can be created by the individual user as required. 

SN reporting also allows for time series trending to be achieved; trending on a daily, weekly, monthly, quarterly and annual basis is straightforward if you are using an existing date/time stamp field, such as the created date or closed date. However, if you’re looking to trend data that changes often, reporting is (for example) unable to plot an open incident in multiple time periods on an axis to show that it’s currently open and has been open in each of the last three months. 

Which types of calculations are available within ServiceNow reporting?

Data aggregation can be applied to reports to create individual calculations, including the following: 

• COUNT: Tell me the total number of incidents created 

• AVERAGE: Tell me the average actual costs related to projects that closed 

• SUM: Tell me the total quantity related to new hardware catalogue requests 

• COUNT DISTINCT: Tell me the distinct number of incidents created 

• MINIMUM/MAXIMUM: Tell me the maximum or minimum processor speed of active computers in our estate 

• STANDARD DEVIATION: Tell me the standard deviation of the time to resolve incident

This is great for individual numbers, but reporting is not able to perform more complex calculations. An example of this is dividing two aggregated numbers by one another, i.e. “tell me the percentage of changes that were successfully closed last month.”

Whilst reports can be created to individually calculate the total number of changes closed and the number that were closed successfully, it is not possible for a report to create the required percentage calculation and divide the two numbers. Without another technology, this type of calculation must be performed through a manual mechanism.

So tell me more about Performance Analytics...

Performance Analytics is an application that’s driven by a defined set of “indicators,” each with clear definition criteria. They are calculated through a scheduled collection job that takes data snapshots of your results on a periodic basis. Think of PA as a tool to help introduce better standardisation and consistency with some of your KPIs and metrics. Each indicator has a definitive calculation and is typically centrally maintained within your organisation by an administrative team. 

An Indicator can be a simple volumetric, such as “number of incidents created” or “number of cases open.” PA also has the ability to reuse metrics within formula indicators, which can enable measurement of KPIs, such as “% of successful changes” and “% of open overdue projects.” These formula indicators will take the results of each of the underpinning indicator scores and then apply the calculation specified within the formula.

An indicator can also have multiple “breakdowns” associated to them. A breakdown is essentially a data dimension that you would like to be able to slice and dice the data by further. Typical examples of breakdowns include: 

• Priority 

• Project Manager 

• Assignee Group 

• Location 

• Company 

• Department 

Without breakdowns, you would be limited to viewing the overall result of the indicator and drill down to the data only. You wouldn’t be able to perform operations such as identifying if a certain location is receiving a lesser level of service in comparison to other areas. Breakdowns become important when we come to look at Analytics Hub and dashboard filtering. 

Each active indicator is included in a scheduled data collection job that typically run daily, but can be changed to other frequencies.  Ultimately each time the jobs are run, the result snapshots are added to a separate “scores” table. 

As we mentioned, a drawback of reporting is that reports can’t look back and trend an incident being open three months ago and plot that same incident against each of those months in a trend report. Within PA, this is solved by taking a daily snapshot of how many incidents were open at the start/end of each day, week or month, allowing you to trend your backlog growth.

The above image shows an example of the scoresheet captured for a “Number of open incidents” indicator. Not only can we see the results of the actual indicator in the top row, but we can also see the numbers that have been collected for each related breakdown.

Now that PA is creating snapshots of results in the scores tables, it’s worth mentioning at this point an important limitation: PA supports a maximum of two breakdowns being applied at the same time.

Let’s imagine that for each day, PA was to attempt to record a full breakdown combination for the below dimensions, with some fairly conservative possible value numbers for each:

Priority (5 records) x Service (10 records) x Location (10 records)

x Department (20 records) x Assignment Group (50 records)

In order for ServiceNow to support the ability to slice and dice any possible combination of the above, it would potentially store 500,000 rows of data just for a single day on one indicator. That’s up to 182.5 million rows of data per year. Whilst some other MI tools may be able to support these volumes, we must remember that ServiceNow is primarily designed for transactional volume as opposed to resource intensive processing and datamining. The limit of a maximum combination of two fields is in place to maintain sustainable load times as well as ensuring it doesn’t negatively impact the overall performance of the platform.

With this limitation now in place, using the example above, the maximum number of records that could be stored for a single day against any indicators using those breakdowns is 1,000 (20 departments x 50 groups). That’s not to say that PA can’t support all of the above breakdowns - we’re just saying that the maximum combination of what you can filter on at any one time is two.

This means that you can filter an indicator for Priority and see that data by service or filter for a specific department and see the result by Assignment Group.  However, if you were attempting to filter on a priority, then want to filter for a certain service, and then see performance broken out by assignment group, this would cause the limit to be reached. When creating your own requirements, it’s important to think about whether this limitation would apply to you for complex calculations.

So that was a hefty one, but we're far from finished...

In Part Two of our Reporting vs Performance Analytics: What's the Difference? series we’ll be going through the importance of Analytics Hub and how this comes into play. We’ll also be looking further into Performance Analytics and its role within calculations. You can read Part Two here.

For more information or a free consultation on how Unifii could potentially help your business, please get in touch with one of our team here.