Dashboard coming up flat

[Jonathan Brown]

Has anyone ever encountered flat graphs in their dashboards? I'm new to Performance Analytics (PA) and dashboarding.

I took over a dashboard after a leadership change a few months ago. I had to update the indicators to represent a specific section in the org chart, not company-wide. I cleared out the scorecards and reran the historical data with new parameters. After each run, I consistently get a flat line for the entire history on dates for "On Hold," "In Progress," and "Resolved." The states for "New" and "Closed" do not share the same issue. For these states, there is a date field for open and closed incidents, which might be the reason. The other states worked prior to my updates, so I'm not sure where I might have gone wrong.

I've been doing this over the last few months and yield the same results. Every reset (60/90 historical) gives the first few weeks like this, and then when new data comes in, it starts to show activity changes. Initially, I thought it might be the data, but would that be the case if every reset shows the same issue?

I could probably live with it and just wait for it to clear up, but after a few weeks, it starts to level out again and flatline. I'm not sure what to check or even look for at this point.

[sizzleMcFace]

Performance Analytics indicators work in two different ways. Some measure the *current state* of records (for example, “How many incidents are in progress today?”) and they record and visualize these snapshots over time, while others measure for example, “How many incidents were created each day?”). The difference is important when it comes to historical data.

When you collect history for an indicator like “State is In Progress,” ServiceNow only knows which records are in that state right now. If you ask it to backfill 90 days, it will apply today’s snapshot to every day in the past, which produces the flat line you are seeing.

In contrast, indicators tied to specific dates, like “Created on” or “Closed on,” can be collected historically, because the system can look back and see exactly when those events happened. When data collection runs for last Monday, "Created on Today" means "Created on field value is last Monday".

That explains why “New” and “Closed” show meaningful history if you don't actually filter for state values but "Created" and "Closed" datetimes, while “In Progress” and “On Hold” do not. Those states don’t have a natural timestamp to replay. The data only becomes useful once you let the scheduled collections run over time. If you need true historical trends for states like “In Progress,” you would need to build indicators based on state transition records (metric instances) rather than just the current condition.