App: VBA Workflow Pro (V3) Scope: Mode 1 (Product Diagnostic) + Mode 3 (User Journey Audit) Date: 2026-04-07 Persona: German office worker (Sachbearbeiter) — knows WENN() formulas, not VBA
Three separate rule builders that do functionally the same thing:
| Feature | Tab Name | Color | Logic Keywords | Saved Store |
|---|---|---|---|---|
| SELECT CASE | xtp-case |
Pink (--pk) |
UND / ODER | S.savedCases |
| IF / ELSE | xtp-ifelse |
Orange (--ora) |
AND / OR | S.savedIE |
| SWITCH | xtp-switch |
Cyan (--cyn) |
AND / OR | S.savedSW |
All three share identical capabilities:
- Same 12 operators (
= != > < >= <= enthaelt beginnt endet leer nicht_leer zwischen) - Same multi-condition support (AND/OR per rule)
- Same per-rule target column override
- Same
{Spaltenname}column reference syntax in result values - Same "first match wins" evaluation strategy
- Same global target + Else fallback pattern
- Same save/load pattern
- Same "Neue Zielspalte" column creation via
prompt()
- The core evaluation engine is solid and correct (condition evaluation, column resolution,
{Spalte}references) - Save/load persists rules to the store for reuse in pipelines
- Per-rule target column is a genuine power feature
- Undo integration (
pushUndo()) before every mutation - Pipeline integration: saved rules can be chained as pipeline steps
- Legacy format backward compatibility in
select-case.js
A) Three tabs, one concept. SELECT CASE, IF/ELSE, and SWITCH are functionally identical. The only differences are cosmetic (color, label, logic keywords DE vs EN). This is the single biggest UX problem. A user building "WENN Land=DE schreibe DAP" has no principled reason to choose one tab over another.
B) No preview. The user clicks "ausführen" and data mutates immediately. There is no dry run, no "what would happen" preview, no before/after comparison. For a data-transformation tool targeting non-programmers, this is high-risk.
C) No visual feedback during building. As the user adds conditions, they see only form elements. No natural-language summary ("If Land equals DE and Betrag is greater than 1000, write Premium in Status"). No match count preview ("14 of 230 rows would match").
D) Mixed-language confusion. SELECT CASE uses UND/ODER (German), IF/ELSE and SWITCH use AND/OR (English). The target audience thinks in German.
E) Value input is blind. The user types a comparison value into a free-text input with no autocomplete from actual column data. If the column "Land" contains "DE", "US", "NG", the user must know this from memory.
F) New column via prompt(). Creating a target column uses window.prompt(), which is jarring and provides no validation or preview.
G) Rule reordering impossible. Rules are appended in creation order. First-match-wins semantics means order matters critically, but there is no drag-and-drop or move-up/move-down.
H) No conflict detection. Two rules can match the same row. The user has no way to see overlapping rules or contradictions.
I) No inline testing. The user cannot pick a single row and ask "which rule would fire for this row?"
J) Saved rules are opaque. The saved rule cards show a terse description (CASE 3 Regeln -> Incoterm). No way to expand, edit, inspect, or clone them.
K) VBA BEAST vs. Workflow Pro: identical code. Both versions of if-else.js and select-case.js are line-for-line identical. No differentiation between the free and pro version in this module.
Step-by-step in SELECT CASE:
| # | Action | Clicks/Inputs | Friction |
|---|---|---|---|
| 1 | Find the TRANSFORMIEREN tab group in the tab bar | 0 (scan) | Must distinguish SELECT CASE from IF/ELSE from SWITCH |
| 2 | Click "SELECT CASE" tab | 1 click | |
| 3 | Select "Standard-Ziel" dropdown (target column) | 1 click + scroll | Must already know target column exists |
| 4 | If target doesn't exist: click "+ Neue Zielspalte" | 1 click | Opens window.prompt() |
| 5 | Type column name in prompt | keyboard | No validation, no preview |
| 6 | Click "OK" on prompt | 1 click | |
| 7 | Click "+ Regel hinzufügen" | 1 click | Adds one case row |
| 8 | In the new row: select source column from "WENN" dropdown | 1 click + scroll | |
| 9 | Select operator (= is default) | 0 if "=" | |
| 10 | Type value "DE" in "Wert" input | keyboard | No autocomplete from data |
| 11 | Confirm target column for this rule (auto-synced) | 0 | |
| 12 | Type result "DAP" in "schreibe" input | keyboard | |
| 13 | Repeat steps 7-12 for each country | N * 6 interactions | Adding 5 countries = 30 more interactions |
| 14 | Optionally type Case Else value | keyboard | |
| 15 | Click "SELECT CASE ausfuehren" | 1 click | Data mutates immediately |
| 16 | See result toast "N rows" | 0 (passive) | No before/after, no row-level detail |
Total for 5 country mappings: ~40 interactions (clicks + key entries), zero visual confirmation before execution.
Comparison: same task in Excel Conditional Formatting: Excel users know: select column -> Home -> Conditional Formatting -> New Rule -> pick values. ~8 clicks for a similar concept, with live preview in cells.
| # | Action | Clicks |
|---|---|---|
| 1 | Navigate to IF/ELSE tab | 1 |
| 2 | Select target column | 1 |
| 3 | Click "+ IF / ElseIf Block" | 1 |
| 4 | Select source column "Land" | 1 |
| 5 | Select operator "=" | 0 (default) |
| 6 | Type "DE" | keyboard |
| 7 | Click "+ Bedingung" | 1 |
| 8 | Select AND/OR (defaults AND) | 0 |
| 9 | Select column "Betrag" | 1 |
| 10 | Select operator ">" | 1 |
| 11 | Type "1000" | keyboard |
| 12 | Type result "Premium" | keyboard |
| 13 | Click "IF/ELSE ausfuehren" | 1 |
Total: ~8 clicks + 3 keyboard entries. This is actually reasonable for a single complex rule. The problem is scale: adding a second ElseIf repeats steps 3-12.
| Feature | VBA Workflow Pro | Excel Cond. Formatting | Notion Formulas | Zapier Filters | Airtable Automations |
|---|---|---|---|---|---|
| Unified rule builder | 3 separate tabs | 1 dialog | 1 formula bar | 1 filter step | 1 trigger config |
| Live preview | None | Yes (cells highlight) | Yes (inline) | Yes (test step) | Yes (test record) |
| Value autocomplete | None | Yes (from data) | Yes | Yes (from schema) | Yes (from field) |
| Natural language | None | None | None | None | None |
| Drag reorder | None | Yes | N/A | Yes | Yes |
| Test single row | None | N/A | Yes (hover) | Yes (test step) | Yes (test record) |
| Conflict detection | None | None | N/A | N/A | N/A |
| Rule templates | None | Built-in presets | Templates | Templates | Templates |
| Undo | Global only | Per-rule | Per-action | N/A | N/A |
Key takeaway: Every competitor with comparable features provides live preview. VBA Workflow Pro is the only tool in this category that mutates data without showing the user what will happen first.
Scoring: Impact (1-5) x Confidence (1-5) / Effort (1-5) = ICE Score
| Rank | Improvement | I | C | E | ICE | Rationale |
|---|---|---|---|---|---|---|
| 1 | a) Merge IF/ELSE + SELECT CASE + SWITCH into one "Regeln" builder | 5 | 5 | 3 | 8.3 | Eliminates the biggest confusion. Three identical engines become one. Reduces codebase by ~60%. Users stop asking "which one do I use?" |
| 2 | c) Live preview row (show match count + first 3 before/after rows) | 5 | 5 | 2 | 12.5 | Highest-value feature missing. Non-programmers need to SEE what will happen. Debounced re-evaluation on every rule change. |
| 3 | g) Smart value suggestions (autocomplete from column data) | 5 | 5 | 2 | 12.5 | Typing "DE" from memory when the column has 12 country codes is error-prone. Show a datalist of unique values from the selected column. |
| 4 | h) Inline row testing ("Test this row" button) | 4 | 5 | 2 | 10.0 | Let user click any table row and see which rule fires, what the result would be. Critical for debugging rules that don't match as expected. |
| 5 | d) Rule templates (Preise-Staffel, Land-Mapping, Status-Workflow) | 4 | 4 | 2 | 8.0 | Pre-built rule sets for common German office scenarios (country -> incoterm, amount -> price tier, status workflow). Reduces blank-canvas paralysis. |
| 6 | e) Drag-and-drop rule reorder | 4 | 5 | 2 | 10.0 | First-match-wins means order is critical. Currently impossible to reorder without deleting and recreating. |
| 7 | j) Rule conflict detection | 4 | 4 | 3 | 5.3 | Show warnings when two rules can match the same row. Prevents silent data errors. |
| 8 | b) Natural language input | 5 | 3 | 4 | 3.8 | "Wenn Land = Nigeria, schreibe Form M erforderlich" parsed into rule. High impact but NLP parsing is effort-heavy and fragile for edge cases. |
| 9 | m) Bulk rule editing (change target column for all rules at once) | 3 | 5 | 1 | 15.0 | Quick win. Already partially implemented via syncTargets() but not exposed as explicit action. |
| 10 | k) Excel WENN() formula import | 4 | 3 | 4 | 3.0 | Parse =WENN(A1="DE";"DAP";"EXW") into rules. High value for migration but formula parsing is complex. |
| 11 | f) Visual rule flow (flowchart/tree view) | 3 | 3 | 4 | 2.3 | Nice visual aid but not critical for the target persona. They think in tables, not flowcharts. |
| 12 | i) Undo per rule | 3 | 4 | 3 | 4.0 | Currently only global undo exists. Per-rule undo would let users experiment more freely. |
| 13 | l) Color-coded rule groups | 2 | 4 | 1 | 8.0 | Quick visual win but low functional impact. |
Re-sorted by ICE score (descending):
| Priority | Improvement | ICE |
|---|---|---|
| P0 | Bulk rule editing (m) | 15.0 |
| P0 | Live preview (c) | 12.5 |
| P0 | Smart value suggestions (g) | 12.5 |
| P1 | Inline row testing (h) | 10.0 |
| P1 | Drag-and-drop reorder (e) | 10.0 |
| P1 | Merge into unified builder (a) | 8.3 |
| P1 | Color-coded groups (l) | 8.0 |
| P1 | Rule templates (d) | 8.0 |
| P2 | Rule conflict detection (j) | 5.3 |
| P2 | Per-rule undo (i) | 4.0 |
| P2 | Natural language input (b) | 3.8 |
| P3 | WENN() formula import (k) | 3.0 |
| P3 | Visual rule flow (f) | 2.3 |
Replace three tabs with ONE "Regeln" tab. Internally, keep one engine (the current evalCond + evalBlock pattern is sound).
New tab structure:
TRANSFORMIEREN:
[Regeln] [Text] [Berechnet] [Formeln] [VLOOKUP]
Instead of three separate files (if-else.js, select-case.js, switch.js), consolidate into one rules.js that:
- Uses German logic keywords consistently (UND/ODER, not AND/OR)
- Exposes one "add rule" button, one "add condition" button
- Keeps per-rule target column override
- Keeps else/default value
- Stores in one
S.savedRulesarray (with migration forsavedCases,savedIE,savedSW)
The existing evalCond() function from if-else.js is already shared by switch.js via import. SELECT CASE has its own _csEvalCond() which is identical code. This is a clear DRY violation begging to be unified.
Add a <div id="rule-preview"> below the rule builder that shows:
- Match count: "23 von 450 Zeilen betroffen"
- First 3 matching rows in a mini-table with before/after columns highlighted
- Debounced: re-evaluate on every dropdown change or value input (200ms debounce)
For inline row testing: clicking any row in the data table while the rule builder is open should highlight that row and show which rule matched (or "keine Regel passt").
- Replace free-text
<input>for condition values with<input list="datalist-{colName}">populated fromnew Set(S.xD.map(r => r[ci])). Max 200 unique values. - Add drag handles to rule rows. Use the existing
drag-drop.jsmodule pattern. On drop, reorder the DOM elements and re-index.
- Shipping templates: Land -> Incoterm, Land -> Zollbehoerde, UN-Klasse -> Verpackungsgruppe
- Finance templates: Betrag -> Preis-Staffel, Konto -> Kategorie
- Conflict detection: after building rules, run all rows through all rules and flag rows where multiple rules match.
The three files share massive duplication:
| Pattern | if-else.js | select-case.js | switch.js |
|---|---|---|---|
| Operator list (12 ops) | _ieOpOpts() |
_csOpOpts() |
_swOpOpts() |
| Condition HTML builder | _ieCondHTML() |
_csCondHTML() |
_swCondHTML() |
| Op change handler (zwischen toggle) | _ieBindOpChange() |
_csCondOpChange() |
_swBindOpChange() |
Column resolver {Name} |
_ieResolve() |
_csResolveResult() |
_swResolve() |
| Condition evaluator | evalCond() |
_csEvalCond() |
reuses evalCond() |
| Block/rule evaluator | evalBlock() |
_csEvalRule() |
_swEvalRule() |
| DOM parser | _ieParseBlock() |
_csParseRules() |
_swParseRule() |
| Add row/block | ieAddElseIf() |
csAddRow() |
swAddRule() |
| Add condition | ieAddCond() |
csAddCond() |
swAddCond() |
| Sync targets | ieSyncTargets() |
csSyncTargets() |
swSyncTargets() |
| New column via prompt | ieAddNew() |
csAddNew() |
swAddNew() |
| Run | IE_RUN() |
CS_RUN() |
SW_RUN() |
| Save | IE_SAVE() |
CS_SAVE() |
SW_SAVE() |
| Render saved | renderSavedIE() |
renderSavedCases() |
renderSavedSW() |
| Run saved | runSavedIE() |
runSavedCase() |
runSavedSW() |
15 duplicated patterns across 3 files = 45 nearly-identical functions that could be 15 functions in one file.
- Do NOT add a fourth rule builder variant. The problem is too many options, not too few.
- Do NOT add a code editor / VBA syntax input. The target user does not write code. Natural language is the stretch goal, not a code editor.
- Do NOT remove the per-rule target column feature. This is genuinely powerful and differentiating vs. simple conditional formatting.
- Do NOT break pipeline integration. Saved rules feed into the pipeline system. Any refactor must preserve
runSavedCase()/runSavedIE()compatibility or migrate the pipeline references. - Do NOT add real-time evaluation on every keystroke for large datasets. Preview should be debounced and limited to first N rows. The app runs 100% client-side; 10k+ rows will cause jank.
- Do NOT localize to English. This is a German-market product. Commit to German throughout (UND/ODER, not AND/OR).
| Metric | Current Baseline | Target |
|---|---|---|
| Tabs needed to build one rule set | 3 options to choose from (confusion) | 1 unified tab |
| Clicks to build a 5-case country mapping | ~40 | ~25 (with autocomplete + templates) |
| Data mutations without preview | 100% (every run) | 0% (preview before every run) |
| Duplicated functions across rule files | 45 | 15 (single-file consolidation) |
| User errors from typos in value input | Unknown (no tracking) | Measurable via autocomplete adoption rate |
| Time to first successful rule (new user) | Unknown | Track via tour/onboarding completion + first rule execution |
The rule builder system in VBA Workflow Pro has a sound evaluation engine underneath a confusing, triplicated UI surface. The core problem is not technical -- evalCond() works correctly -- but structural: three tabs that do the same thing, no preview before mutation, and no assistance during value entry.
The highest-leverage changes are:
- Merge three builders into one "Regeln" tab (eliminate confusion, cut code by 60%)
- Add live preview (show match count + sample rows before execution)
- Add value autocomplete from column data (prevent typos, accelerate input)
These three changes would transform the rule builders from a power-user tool that requires VBA mental models into an accessible data-transformation interface that matches the Excel-familiar expectations of the target persona.