Sap Signavio

SAP Signavio Process Management & Mining

Related Skills

• sap-build-process-automation — Automate processes modeled in Signavio
• sap-s4hana-extensibility — Process variants for S/4HANA scenarios
• sap-cloud-alm — ALM process monitoring complements Signavio insights

Quick Start

Choose your Signavio component:

First API call — List Process Models:

curl -X GET "https://editor.signavio.com/p/revision" \
  -H "x-signavio-id: {{SESSION_TOKEN}}" \
  -H "Accept: application/json" \
  -b "JSESSIONID={{SESSION_COOKIE}}"

Core Concepts

BPMN 2.0 Modeling Standards

Process Intelligence (Mining) Pipeline

1. Data extraction — Pull event logs from source systems (S/4HANA, ECC, SuccessFactors)
2. Data transformation — Map to Signavio event log format (Case ID, Activity, Timestamp)
3. Process discovery — Auto-generate process model from event data
4. Conformance checking — Compare discovered vs. designed process
5. Root cause analysis — Identify bottlenecks, rework loops, deviations
6. KPI monitoring — Track cycle time, throughput, automation rate

Event Log Format

CaseID,Activity,Timestamp,Resource,Variant
PO-001,Create Purchase Requisition,2026-01-15T08:00:00Z,User_A,Standard
PO-001,Approve Purchase Requisition,2026-01-15T09:30:00Z,Manager_B,Standard
PO-001,Create Purchase Order,2026-01-15T10:00:00Z,System,Standard
PO-001,Goods Receipt,2026-01-20T14:00:00Z,User_C,Standard
PO-001,Invoice Receipt,2026-01-22T11:00:00Z,User_D,Standard
PO-001,Payment,2026-01-25T16:00:00Z,System,Standard

Common Patterns

Pattern 1: Signavio API Authentication

import requests

SIGNAVIO_URL = "https://editor.signavio.com"

def authenticate(email, password, tenant_id):
    """Authenticate and get session token."""
    resp = requests.post(
        f"{SIGNAVIO_URL}/p/login",
        data={
            "name": email,
            "password": password,
            "tenant": tenant_id,
            "tokenonly": "true"
        }
    )
    resp.raise_for_status()
    return {
        "token": resp.text.strip(),
        "cookie": resp.cookies.get("JSESSIONID")
    }

def get_headers(session):
    return {
        "x-signavio-id": session["token"],
        "Accept": "application/json",
        "Cookie": f"JSESSIONID={session['cookie']}"
    }

Pattern 2: Export BPMN Model

def export_bpmn_xml(session, model_id):
    """Export process model as BPMN 2.0 XML."""
    headers = get_headers(session)
    resp = requests.get(
        f"{SIGNAVIO_URL}/p/revision/{model_id}/bpmn2_0_xml",
        headers=headers
    )
    resp.raise_for_status()
    return resp.text

def export_svg(session, model_id):
    """Export process diagram as SVG."""
    headers = get_headers(session)
    headers["Accept"] = "image/svg+xml"
    resp = requests.get(
        f"{SIGNAVIO_URL}/p/revision/{model_id}/svg",
        headers=headers
    )
    resp.raise_for_status()
    return resp.text

Pattern 3: Process Mining Data Preparation (S/4HANA P2P)

-- CDS view for Procure-to-Pay event log extraction
@AbapCatalog.sqlViewName: 'ZV_P2P_EVLOG'
@Analytics: { dataCategory: #FACT }
define view Z_P2P_EVENT_LOG as

  -- Purchase Requisition Created
  select from eban {
    cast(banfn as abap.char(20)) as CaseID,
    'Create Purchase Requisition' as Activity,
    badat as EventTimestamp,
    ernam as Resource,
    bsart as Variant
  }
  union all
  -- Purchase Order Created
  select from ekko {
    cast(ebeln as abap.char(20)) as CaseID,
    'Create Purchase Order' as Activity,
    aedat as EventTimestamp,
    ernam as Resource,
    bsart as Variant
  }
  union all
  -- Goods Receipt
  select from mkpf
    inner join mseg on mkpf.mblnr = mseg.mblnr
                   and mkpf.mjahr = mseg.mjahr {
    cast(mseg.ebeln as abap.char(20)) as CaseID,
    'Goods Receipt' as Activity,
    mkpf.budat as EventTimestamp,
    mkpf.usnam as Resource,
    mseg.bwart as Variant
  }
  where mseg.bwart = '101'

Pattern 4: BPMN Best Practices Template

<!-- Well-structured BPMN process -->
<bpmn:process id="P2P_Standard" name="Procure-to-Pay Standard" isExecutable="true">
  <!-- Single start event -->
  <bpmn:startEvent id="Start" name="Purchase Need Identified"/>

  <!-- Clear task naming: Verb + Object -->
  <bpmn:userTask id="T1" name="Create Purchase Requisition"/>
  <bpmn:userTask id="T2" name="Approve Purchase Requisition"/>

  <!-- Gateway with labeled outgoing flows -->
  <bpmn:exclusiveGateway id="G1" name="Approval Decision"/>
  <bpmn:sequenceFlow id="F_Approved" name="Approved" sourceRef="G1" targetRef="T3"/>
  <bpmn:sequenceFlow id="F_Rejected" name="Rejected" sourceRef="G1" targetRef="T_Reject"/>

  <bpmn:serviceTask id="T3" name="Create Purchase Order"/>
  <bpmn:userTask id="T4" name="Confirm Goods Receipt"/>
  <bpmn:serviceTask id="T5" name="Process Invoice"/>
  <bpmn:serviceTask id="T6" name="Execute Payment"/>

  <bpmn:endEvent id="End_Success" name="Payment Completed"/>
  <bpmn:endEvent id="End_Rejected" name="Requisition Rejected"/>
</bpmn:process>

Pattern 5: Process KPI Configuration

{
  "kpis": [
    {
      "name": "Cycle Time",
      "description": "Average time from requisition to payment",
      "formula": "AVG(EndTimestamp - StartTimestamp)",
      "target": "< 15 days",
      "dimension": "days"
    },
    {
      "name": "First-Time-Right Rate",
      "description": "Percentage of cases without rework loops",
      "formula": "COUNT(cases_no_rework) / COUNT(all_cases) * 100",
      "target": "> 85%",
      "dimension": "percentage"
    },
    {
      "name": "Automation Rate",
      "description": "Activities executed by system vs. total",
      "formula": "COUNT(system_activities) / COUNT(all_activities) * 100",
      "target": "> 60%",
      "dimension": "percentage"
    },
    {
      "name": "Maverick Buying Rate",
      "description": "POs created without prior requisition",
      "formula": "COUNT(po_without_pr) / COUNT(all_po) * 100",
      "target": "< 5%",
      "dimension": "percentage"
    }
  ]
}

Error Catalog

Performance Tips

1. Event log size — Keep under 10M events per analysis; sample or partition by date range
2. Activity normalization — Standardize naming before import; "Create PO" and "PO Creation" = same activity
3. Case linking — Use document flow tables (VBFA, EKBE) to link cross-module cases
4. Incremental mining — Schedule delta extracts (changed since last run) for ongoing monitoring
5. Model complexity — Keep BPMN models under 50 elements per diagram; use subprocesses for detail
6. Variant analysis — Focus on top 20 variants (usually 80%+ of cases); group rare variants as "Other"
7. Collaboration Hub — Publish only approved models; use governance workflow for version control

Gotchas

• Signavio vs. BPA: Signavio models are design-time reference; SAP Build Process Automation is runtime execution — they complement, not replace each other
• BPMN compliance: Signavio enforces strict BPMN 2.0; models from other tools may need cleanup before import
• Event timestamp granularity: Second-level timestamps required for accurate mining; day-level creates artificial parallelism
• Process Intelligence licensing: Mining capabilities require separate PI license; Process Manager alone doesn't include mining
• S/4HANA connector: Pre-built extractors exist for O2C, P2P, AP — custom processes need manual event log setup