8 Advanced Performance Analysis

This chapter sharpens the dashboard you built in Chapter 7 into a decision tool. The aim isn’t more charts; it’s better judgment. We’ll read the pulse of throughput and service against targets, compare performance across periods so we don’t chase noise, and use a Pareto to focus on the few issues that cause most of the pain. You’ll see how to frame a finding (“what changed, by how much, and where?”), and how to turn it into a concrete action for the coming week. Everything runs in Power BI using the datasets you already loaded.

8.1 What “advanced” means here

Advanced performance work is not about more sophisticated formulas; it’s about separating signal from noise so managers intervene where it matters. In operations, random day-to-day bumps are normal. We care about runs (several days below target), shifts (a new level), and where the misses concentrate (by region, route, line, or blend). The rest of this chapter teaches you to see those patterns quickly and defend your interpretation with simple, honest visuals.

8.2 The pulse: throughput and on-time vs. target

Start with the pulse: two compact visuals that answer “Are we on pace?” and “Are we meeting service?” Place On-Time % as a daily line filtered to the last 28 days, and a companion visual for Throughput Units over the same window. Add the target as a constant line (Analytics pane) so the eye knows immediately whether we’re where we should be.

How to read the pulse:

• Look for runs below the target (e.g., two or more consecutive days). That’s an operational event, not random noise.

• Scan for day-of-week rhythm. If Fridays and Saturdays dip, that’s a capacity or sequencing issue during a predictable peak.

• Pair the story: if throughput spikes on promo days while on-time drops, the likely cause is load, not a carrier failure.

Make the visuals carry the story. Titles state the takeaway (“On-time below 95% for two of the last five days”), axes include units/percent, and one short annotation calls out an exception (“Promo 8/15”). If labels crowd, reduce detail rather than shrinking text—managers should be able to read this on a laptop or projector without squinting.

8.3 Rolling comparisons: week-over-week and month-over-month

A pulse can look alarming until you compare it with the prior period. Rolling comparisons keep us from overreacting to seasonality.

Use week-over-week for daily dashboards and month-over-month for monthly reviews. In Power BI, you can duplicate the pulse visuals and switch the page’s date range to the previous period, or add a paired card showing the change (e.g., On-Time % Δ vs last week) using Quick Measures → Time intelligence (percentage difference from previous period). The important part is interpretation: are we trending up or down relative to a comparable window, and is that movement operationally meaningful?

Reading the comparison:

• If On-Time % is down vs last week and the deficit clusters on the same weekdays, the fix is likely staffing or sequencing rather than a random blip.

• If throughput is higher but service held steady, celebrate: that’s evidence your process absorbed a surge.

• If both worsened, link back to the slice that explains it (region, blend, route, or line).

Keep the display simple: one Δ card per KPI, green when better than last period, red when worse. You’re teaching managers to expect context, not just a snapshot.

(Clicks only: Quick Measure → Time intelligence → “Month over month change” or “Previous period” requires marking Calendar as the Date table, which you already did in Ch. 7.)

8.4 Pareto: focus on the biggest few

Once you know performance moved, the next question is where to intervene. The Pareto principle says a small number of categories often account for most of the problem. Build a sorted bar chart (descending) of your chosen driver—defect type, region, route, or line—using the adverse metric (Defects, or Late Units). Add a Top N filter (e.g., Top 5) so the eye lands on the levers first. If you’d like to emphasize proportion, show data labels as percentages of the total.

How to read it:

• Name the concentration: “Two routes account for 57% of late units this month.”

• Ask for plausibility: does this match what supervisors see on the floor or in the field?

• Propose the smallest action that could move the needle (reroute, add a shift, recheck a line).

The aim isn’t a perfect 80/20 curve; it’s a ranked list that credibly points to where action will help most.

8.5 Stability check: simple rules without full SPC

You don’t need a full control chart to avoid knee-jerk reactions. Two simple heuristics catch most meaningful shifts:

1. Two-day rule: two or more consecutive days below target merits a look.

2. Four-of-five rule: if four of five recent days are below target—even if not consecutive—flag it.

Use these as review triggers, not conclusions. Triggers send you back to the slice or process step that explains the miss. On the page, a small text box labeled “Triggers this week” lists which rule fired and on which dates. That makes your thresholding explicit and repeatable

8.6 Applied assignment — Performance Pulse (Power BI)

What you submit (lightweight):

• One Power BI page that extends your Chapter 7 dashboard with:

  1. a 28-day On-Time % pulse (target line, one annotation),

  2. a Throughput Units pulse aligned to the same dates, and

  3. a Pareto (sorted bars with Top N) for the driver you choose (defect type, region, route, or line).
     Add two small Δ cards (week-over-week or month-over-month) for On-Time % and Throughput.

• One paragraph memo (5–6 sentences) that states the question, summarizes the pulse and the comparison, names the top contributor from the Pareto, and commits to one action with owner and timing.

How it’s graded:

• The page makes the status clear (targets visible, Δ cards readable, labels meaningful).

• The Pareto is sorted and focused (Top N).

• The memo proposes a realistic action for the next week.

Chapter 8 References

American Society for Quality. (n.d.). What is a Pareto chart? Analysis & diagram. ASQ. https://asq.org/quality-resources/pareto

Microsoft. (2024, March 15). Use quick measures for common and powerful calculations. Microsoft Learn. https://learn.microsoft.com/en-us/power-bi/transform-model/desktop-quick-measures

Microsoft. (n.d.). What is Power BI? Microsoft. https://learn.microsoft.com/en-us/power-bi/fundamentals/power-bi-overview

OpenStax. (2019). Principles of management. OpenStax, Rice University. https://openstax.org/books/principles-management/pages/6-introduction