KPI - Process Performance Assessment: Every Property, Every Department, One Scorecard
The first six posts in this series made one point from six angles: what breaks a measurement system is usually not the data, it is the structure. Without a goal tree, an indicator does not know what it serves. Unwritten weights get relitigated every month. Frequency held at scorecard level instead of per indicator produces false alarms in July. A wrong scoring direction rewards the months that got worse.
Each is solvable on its own — in a spreadsheet, even. Holding all six at once, in one place, is the hard part.
This post is how HRMIM does that. The product is called KPI - Process Performance Assessment. I will describe what it does and what it does not do; the second half is usually more useful.
Where the previous six posts left off
Each post produced a requirement.
| Topic | The requirement it produced |
|---|---|
| Goal tree | Four layers, one structure |
| Construct validity | Definitions in one place, reviewable |
| Weighting | Weight able to differ by property |
| Measurement frequency | Frequency defined per indicator |
| Target direction | Each indicator carries its own scoring method |
| Outliers and comparison | Deviation and periods in the same report set |
Six requirements, one condition: definition, measurement and scoring must sit in the same structure. In separate files they get aligned by hand every month — until the month they do not.
What measuring in one structure means: the property × department breakdown
KPI - Process Performance Assessment evaluates company performance in a property × department breakdown, monthly. Front Office and Housekeeping at one property are scored separately; property and group totals roll up from there.
Evaluation is periodic: a calendar year maps to one period, and each property runs its own. That is what lets properties on different opening dates or seasonal calendars coexist in one group.
"The group score dropped" means nothing on its own. In a property and department breakdown, where the drop came from becomes a reading rather than an argument.
The four-layer goal tree: Objective → Goal → Sub-Goal → Indicator
The spine is four layers: Amaç → Hedef → Alt Hedef → Gösterge (Objective → Goal → Sub-Goal → Indicator). The objective carries the group's intent, the goal its property-level equivalent, the sub-goal departmental ownership, the indicator the single number measured.
Same logic as the goal tree framework in the first post. The difference: there it was a document, here it is the data model. Open an indicator and its sub-goal, goal and objective come with it, always.
The four layers are Kaplan and Norton's 1992 scorecard framework brought down into a multi-property structure.
Weighting: indicator and department, able to differ by property
Every indicator carries an indicator weight, every department a department weight. Not every indicator is equal — the critical ones carry more.
The part that matters: weights can differ by property. An indicator can be more critical at one hotel than another, and the system records that as a definition, not an inconsistency. Everything we argued about the financial logic of weighting depends on it.
Measurement frequency: monthly to annual, per indicator
Each indicator is configured with how often it is measured in a year: monthly, every two months, quarterly, every four months, half-yearly or annually.
Frequency belongs to the indicator, not the scorecard — so an annual brand standard audit and monthly occupancy sit side by side without distorting each other.
Telling missing data apart from not-yet-due data
The payoff of per-indicator frequency shows up in the report. An indicator that should have been measured and was not entered is never confused with one that was not due this month.
The distinction looks small and is not. Without the separation between missing and not-yet-due data, a manager chases blank cells every month and misses the ones genuinely absent.
Five scoring methods
An indicator can be scored against its target five ways.
| Method | When it is the right one |
|---|---|
| Actual / Target | Higher-is-better indicators with no natural ceiling |
| Target / Actual | Ratio readings where the actual never approaches zero |
| Higher than target | Cases where the target is a threshold |
| Lower than target | Lower-is-better indicators |
| Closeness to target | Indicators where deviation either way is a loss |
More is not better for every indicator. Revenue rising is good, turnover rising is bad, stock days should sit close to target — one formula cannot score all three. Target direction and scoring methods has the arithmetic; the product binds the method to the definition.
An indicator with no target is not scored
If no target has been entered, the indicator is not scored. Not quietly zero, not quietly a hundred.
A deliberate data-integrity decision, not a gap. Zero reports something unmeasured as a failure; a hundred inflates the scorecard with an unverified number. The system leaves the indicator out of scoring and makes the gap visible — the decision stays with you.
The report set: 10 pages, which report answers which question
Ten pages: Dashboard, Data Entries (Veri Girişleri), Data Control (Veri Kontrol), Overall Results (Genel Sonuçlar), Results Analysis (Sonuçlar Analizi), Outliers (Aykırılıklar), Strategic Analysis (Stratejik Analiz), Comparison (Karşılaştırma), Segmentation (Segmentasyon), Analysis Tree (Analiz Ağacı). Every screen exports to Excel. Five of them carry the month-end cycle.
Data Control (Veri Kontrol)
Which cell was filled, which is missing. The audit screen for month-end close, where you stop before the scorecard goes out.
Outliers (Aykırılıklar)
Surfaces indicators far below or far above target. An extreme deviation is not automatically bad performance — it can also mean the target was set wrong. The decision matrix in the outlier analysis post describes how to read it.
Comparison (Karşılaştırma)
Period over period — 2025 against 2026 — along property, department, category, objective, goal, sub-goal or indicator. Comparison only means something while definitions and weights hold still; the conditions for valid cross-property comparison apply here too.
Segmentation (Segmentasyon)
Breaks the total into its components. First stop for "what pulled this score down".
Analysis Tree (Analiz Ağacı)
Follows the same question through the goal tree layers, from indicator back up to objective.
Zek.ai: narrative commentary on a completed period
Zek.ai, the platform's AI layer, reads the results once a period completes and produces a narrative commentary: the overall picture, the standouts, the weak areas.
Worth being precise. Zek.ai writes a commentary; it does not decide. Which target gets revised, which deviation is operational and which definitional — those stay management calls. The AI layer shortens the reading in front of the decision, not the decision itself.
The monthly cycle, from period open to close
Measurements are entered centrally. The Gösterge İşlemleri screen is an Excel-like grid you move through by keyboard; measurement and note fields are filled in the cell.
You can work department by department or indicator by indicator, and the completion percentage is visible on screen, so what is outstanding takes no hunting. Then you close out in Data Control, the scoring engine computes the result, and reports come out from the same definitions.
Be plain about one thing: the product does not collect the data. Measurements get entered. The gain is that entered data is bound to one definition set and the arithmetic is no longer done by hand.
Decisions to make during setup
None of the setup decisions get made for you:
- Which indicators are group-mandatory and which are left to the property
- Each indicator's target, unit and owner
- Indicator and department weights, per property
- Each indicator's measurement frequency
- Each indicator's scoring method
Those decisions are your measurement system, and no structure repairs them once made badly. What lightens the load is that targets and weights can be worked through an Excel template rather than typed cell by cell.
Using the industry's groundwork helps too. USALI standardises revenue and cost lines by department; you add your operational and HR indicators on top.
Who this is for
KPI - Process Performance Assessment was built for companies running more than one property or branch that have to measure them comparably. In a single-property operation the breakdown buys you less.
It runs inside the HRMIM platform. Company- and property-level authorisation, TR/EN multi-language and shared reporting infrastructure with the other HRMIM products come as standard.
There is an international standard for structured human-capital reporting too: ISO 30414:2025 — Human resource management: Requirements and recommendations for human capital reporting and disclosure, now in its second edition. Context, not a compliance claim — defined, owned, periodic indicators are where the sector is heading.
Conclusion
All the discipline described across the six posts can be applied by hand. A careful team with a good file set will do it. The problem is not care, it is endurance: repeating it every month, for every property and department, wears down over a year.
What KPI - Process Performance Assessment does is move that repetition into the structure: four layers built once, weights defined per property, frequency written onto the indicator, the scoring method part of the definition, one report set from one set of definitions.
What it does not do is equally clear. It does not collect your data, set your targets, tell you which weight is right, or decide for you. It does not score an indicator with no target, and Zek.ai comments without concluding. The quality of the system is capped by the decisions you make at setup — true of every measurement system, and we will not claim otherwise.
The gain sits here: the month-end meeting stops being about whether the number is right and starts being about what the number means.
If you carry measurement responsibility across a multi-property group and want to work out which part of your scorecard is a structure problem and which part a definition problem, let us set up a session. To see KPI - Process Performance Assessment against your own indicator set, get in touch and we will start there.


