Signavio carves out an interesting niche – empowering the masses to accelerate transformation efforts through a combination of process modelling (BPMN) and decision modelling (DMN), along with a set of lightweight dictionary, workflow and publishing components. It’s a compelling cloud-based combination for major corporations. Putting process and decision modelling on steroids, the vendor has amassed some pretty big name clients, with an impressive number of licensed decision modelling users. We take a close look at their decision modelling capabilities and how it fits within their suite of tools.
The Signavio Business Transformation Suite is a cloud-based collection of repository-oriented modelling tools to capture, design and publish both processes and decisions, along with the associated goals, strategies and business capabilities, in support of enterprise change initiatives. It also provides internal facilities to simulate, animate and execute those processes and decision models, or export them for execution in other environments. We reviewed Signavio Process Manager 11.0.0 “Enterprise Plus” edition. There are four core components to the product set:
- Process Manager: provides tooling for modelling business processes (BPMN 2.0), decisions (DMN 1.1), as well as tools for capability modelling (via Archimate), simulation and variance management.
- Collaboration Hub: a publishing environment for models built using Process Manager providing both model navigation as well as facilitates for users to comment and make suggestions.
- Workflow Accelerator: a process execution environment that enables flexible governance of the models and the overall environment, as well as automation of simple people-centric processes.
- Process Intelligence: a process mining and analysis component, enabling firms to reverse-engineer processes from system logs, and carry out variance analysis between design and implementation.
Signavio embeds Decision Management capabilities into these components rather than selling a separate modelling tool. The decision modelling element is now only available with the Enterprise Plus edition. This is a relatively new development as the firm had previously sold a standalone modelling tool for decision management. Similarly, all these tools leverage the Dictionary functionality in different ways.
The role of Decision Management
The Signavio Suite integrates decision management into the relevant areas of the core components. The Process Manager component provides the environment to model and design decisions, while also integrating decision-making with process modelling. The Workflow Accelerator uses decision models when working out sophisticated routing for work items. Finally, the Collaboration Hub provides mechanisms for publishing decision models to the enterprise and crowdsourcing ideas for their improvement. The product supports the export of decision models using the DMN XML standard. Signavio supports conformance to level 2 of the standard, and some aspects of conformance at level 3.
The Signavio Suite facilitates business analysts and subject matter experts in developing a reusable set of business processes, decision models and organisational change artefacts. From its inception, Signavio has focused on enabling or empowering “the masses”, without the need for process or decision experts. The product also sets out to enable the broader business community to contribute to and collaborate around improvement initiatives. Through this wider business engagement and by enabling co-operation on change efforts, Signavio seeks to drive insights amongst its users and increase the velocity of business transformation programmes.
From a Decision Management point of view, the product set provides a unique feature that we have not seen elsewhere. Rather than assuming that all decision models lead to execution, Signavio provide a “Decision Assistant” to explore how even partial matches of the input data can lead to differences in the decision outputs. You could think of this as a simple default user interface for the decision itself, enabling a user to execute the model for the data at hand.
Given that the overall product set competes with Visio and PowerPoint at one end of the scale, and enterprise architecture repository-oriented tools at the other, Signavio has staked out an interesting modelling niche. For example, the inclusion of the Workflow Accelerator gives it governance capabilities not available from any of its competitors. Similarly, integration of process mining and variance analysis functionality adds a new dimension to the usage scenarios possible with these sorts of tools.
At present, the Decision Assistant feature is limited to a) testing for decision models under development within the Process Manager component, and b) in the Collaboration Hub where business users can validate and comment on the developed models. However, we feel this underplays its potential of the functionality. Given that the real world is full of all sorts of “what if” scenarios, there’s also a possibility of using Decision Assistant as a lightweight mechanism for execution of standalone decision models. For example, a young driver seeking car insurance might want to experiment with options on choice of car and excess premium. Similarly, a financial advisor might want to explore return investment options based on client circumstances, propensity for risk and history.
Of course, there are always improvement opportunities around the user interface and help systems. For example, the Dictionary tooling is a little simplistic and hard to understand with regard to complex data structures. Yes, one can apparently integrate it with external data sources such as the organisation’s existing EA repository tool, but this requires work at the API level rather than configuration in the tooling. Finally, we see great opportunities for the Workflow Accelerator to support a set of “wizards” that automate aspects of model development, deployment and governance.
Signavio focuses on helping organisations transform their businesses, by combining the ability to leverage the wisdom of the crowd while generating actionable insight from existing systems. It provides an integrated environment for capturing business decision logic and linking that easily to process models, capability models and the strategy models that inform them. Rounding out the suite are the publishing environment and workflow componentry.
Signavio has recognised that, quite apart from automated execution, there exists a very wide variety of use cases for decision modelling. Just like process modelling, many decision management challenges are about visualising and then simplifying business practices.
Decision modelling capabilities come as part the Enterprise Plus edition of the Process Manager. To publish static models beyond the immediate modelling team implies purchasing Collaboration Hub licences. Adding the governance capabilities means additional Workflow Accelerator licenses. While a user can sign up for access for a free 30-day trial of the Process Manager, enterprise licensing is based on the number of modellers, the number of concurrent users accessing the Collaboration Hub, and the size of the enterprise. Given that Enterprise Plus edition targets larger corporates or those dealing with complex challenges, customers typically spend in the range of $20,000 to over $100,000 per annum.
Signavio claims that there are over one million people consuming the models created by 100,000 users of the Process Modeller within more than 1,000 companies. Furthermore, with 6,000 licensed users of the Enterprise Plus configuration, this represents a dominant market share for DMN modelling licenses. It is hard to tell how many of those licensees are active DMN users.
Signavio is emerging as a significant vendor in the modelling repository space, with annual revenues around $20 million in recurring SaaS annual licenses. Annual growth rates are running at over 100% per annum. The firm took an investment of €31 million to fuel growth in December 2015.
Cameo Case Studies
Case Study: A “Big Four” consulting practice uses decision modelling to streamline accounting
All companies carry out transactions of some sort – a derivatives trade, a mortgage, a car sale, even purchase of materials on a construction job. The ever-evolving laws of the land, and the choices or needs of the customer govern the transaction. Yet, after the fact, there are many different and unique reasons to look at the transaction. Each reference to the data set has costs associated with it, and no new value for the original customer.
Consider this scenario. First, a trade goes into the management accounts, in compliance with the business practices, rules, accounting regulations and policies that apply. The initial job-to-be-done here is to create an accurate record of sales and calculate direct sales costs. Of course, there might be bonus or commission to pay, which also affects personal taxation. Then there is the monthly VAT calculation, and quarterly reporting to the markets. On an annual basis, that data goes into the financial statements and tax returns to calculate the corporate liability. Before you know it, along come the auditors. Moreover, it might not stop there – a later tax inspection … leading to a criminal fraud case.
All these uses cases reference the same source data; but with different constraints, rules and objectives. In each scenario, after the transaction occurred, expensive specialist resources (humans) apply their judgement and make decisions within their parochial context.
That set of scenarios represents the challenge that this consulting practice focuses upon. By modelling these different sets of decisions in DMN (and sometimes combining them with BPMN), it becomes possible to factor all of these different nuances into whether that transaction should even happen in the first place. It also enables the consultancy to better industrialise and scale its own IP and knowledge. Moreover, the consultancy is finding that it can provide better support to clients in transforming their business operations with decision modelling as a core competence. Of course, enterprise customers can more easily avoid costly compliance issues, and they get a dramatic reduction in the costs associated with maintaining management accounts and undertaking audits.
Case Study: Telco supports next best action advice to user
This European telco was one of the first users of Signavio products, taking out an enterprise-wide license. Signavio is available to every employee, just as you would expect to find for Microsoft Word or an email client: it currently has 2,200 employees as registered users of Signavio, all of whom have access to decision modelling. It’s no surprise that process modelling still dominates their prime usage, with the addition of decision modelling being a relatively recent addition to the suite.
The primary use cases for decision modelling are around the customer service arena and the “offer-to-order” process. In customer service, the same set of decision models guide call centre agents and self-service interfaces for diagnosing and fixing internet connection and configuration issues. Every question gathers another element of data for input into the decision logic, which in turn opens up new avenues for exploration and assessment. In the offer-to-sales process, decision models help guide employees and clients in customising their purchase to more closely match their needs.
As with any change in operating method, there are challenges. For example, there is a certain degree of method needed to help business users translate their excel spreadsheets, overloaded with different flavours of business rules, into clear and articulate decision logic. This is a common challenge for vendors and is not limited to Signavio.
The Signavio Suite splits into four logically distinct components as shown in the figure below:
- Process Manager provides the ability to model and capture current/future state processes and decisions. Business analysts and business subject matter experts (SMEs) develop their process models using BPMN 2. Alongside that, DMN 1.1 decision models are developed and link directly to the processes to which they relate. Process Manager also provides mechanisms to enable users to model business capabilities (using the Archimate standard), simulate processes (using discrete event simulation) and manage variance.
- Collaboration Hub acts as an enterprise-wide knowledge sharing facility. It provides a sort of enterprise-wide equivalent of a “water cooler” where employees can highlight process and decision issues. It handles processes and everything related to them—decisions, customer journeys and other organisational information—enabling employees to consume that information in a controlled environment, and capturing related feedback. Users can highlight problems and/or suggest improvements.
- Workflow Accelerator offers a lightweight process orchestration layer. This is not intended as an application development platform as with most major BPM Suites. Rather, it sets out to provide a quick and simple method of executing processes contained within the repository—managing business artefacts and attachments. The Workflow Accelerator can use decision models to support complex routing and exception handling scenarios.
- Process Intelligence delivers an automated process discovery/mining tool. This allows customers to leverage data that is already available in the organisation—the system logs and audit trails of existing systems. Process mining involves identifying patterns in this historical data and deriving the actual processes in use, rather than the perceived model developed by analysts. Of course, this only works for processes that are already automated in some way, and doesn’t do much to help with manual steps in those processes. The Process Intelligence tool enables a gap analysis between expected models and observed behaviour, exposing these insights through real time process dashboards.
Given that our core focus in this set of reviews is on the decision management aspects, we have concentrated our review on the Process Manager. Rather than reviewing the entire Signavio Suite, we only tangentially explore how other components leverage decision management functionality.
All the Signavio components are web-based. Users only need a web browser to develop modelling content in Process Manager, or to explore existing content in the Collaboration Hub. The “editor” components provide the user interface to edit decision (DMN) and process (BPMN) models.
Users add new shapes via drag and drop facilities, or in the case of process models via the text/table based Quick Model capability. They move shapes around to suit their needs utilizing a set of utilities to arrange and align them. The attribute panel slides out of the right hand side of the user interface to modify all attributes of the selected shape. As shown in the figures below, double clicking a decision shape within a Decision Requirements Diagram (DRD) opens the decision logic editor.
In the figure above, we are looking at how the Eligibility Scoring is calculated. You’ll notice that the hit policy is set C+, returning the sum of eligibility scores for all rows that return true when assessed. The tool supports all the relevant hit policies of DMN.
Input Data is either a) defined inside the model itself, or b) linked to a Dictionary item. Decision Logic then references the input data and modelled sub-decisions to provide an output.
The Expression Editor supports S-FEEL (see above) and boxed contexts. S-FEEL is the simplified version of the Friendly Enough Expression Language defined within the DMN specification. However, expressions cannot directly access Dictionary elements – users will need to create Input Data Items that map to the Dictionary.
Modelling conventions are fully customizable. For example, customers can ensure the absence, or presence, of certain shapes; or the presence of names and descriptions for shapes, or even the consistent use of Dictionary elements. Customers can define and activate modelling conventions that are checked on save and/or approval.
Signavio delivers the entire Signavio Suite as a cloud based service. At the back end, Signavio uses AWS to store the data and can provide options to ensure that data stays within a particular country or region. Under the hood a repository/database environment manages access to all elements in a client repository. While Signavio recommends the SaaS solution, an on premise solution is also available. It’s relatively trivial to transfer modelling content from SaaS to on premise and/or vice versa. Customers can register online and test the software free of charge for 30 days.
To develop models, users access the browser-based Process Manager component to create models in DMN and BPMN. The Dictionary provides a basic mechanism for modelling the business language of the domain to enable consistent reuse across models.
In the Signavio tooling there are two options available.
- Define everything within a specific model, including all data types of Input Data. Such “locally defined” data types are not reusable by other models.
- Use the “Dictionary” component to define, organize, and manage all terms and data types. This is the recommended approach.
Dictionary items ensure the consistent use of commonly agreed terms and data types and are reusable across many models. The Dictionary is a separate component as shown below.
It’s also possible to populate and maintain Dictionary content from external applications via REST APIs. Subtly different, users can load specific Dictionary items dynamically from external sources – for example, the allowable values for enumerations (lists) and external taxonomies.
The Dictionary does make it possible to model complex entities using hierarchies, but all the models we reviewed used relatively simple name descriptions.
As shown in Figure 2, the DMN modelling canvas provides direct support for standard Decision Requirements diagrams. Figure 3 shows the Decision Logic editor, which also allows customers to verify Decision Tables for completeness and rule overlaps (see below). It shows and highlights the rules that overlap, including the affected value ranges (gaps) not covered by any rules.
One area where Signavio strays from the current DMN standard is in the area of multi-instance decisions (in much the same way that BPMN 2.0 has multi-instance sub-processes). This is seen as a pragmatic mechanism to overcome one of the recognised limitations and challenges in the DMN standard. Modellers need the ability to apply complex decision logic to elements in lists and tables – i.e. multiple connected decisions, applied to each element of the list. Signavio has also inserted list operators directly into decision tables. While this sort of extension may solve problems in the short term for modellers, longer term, models may need tweaking as the OMG updates the standard to solve for these sorts of challenges. In addition, it means that any execution engine used would also need to recognise these non-standard extensions.
Related modelling areas
Since DMN is part of the wider Signavio Process Modelling solution, it’s relatively trivial to integrate decision models into related process models. Typically, task elements of BPMN process models are set to “Rule Tasks” and then linked to a decision model. The repository keeps tracks of all processes models that linked to a specific decision model.
Signavio also supports the Archimate Enterprise Architecture modelling standard within the Enterprise Plus version of the modelling suite. Amongst other modelling canvases, this provides the ability to model business capabilities and business strategy.
Decision Management interfaces
Two main options are available:
- Decision models are exported via the standardised DMN XML format and the executed in DMN engines such as those provided by Camunda or Red Hat.
- Signavio can generate DRL code for a selected decision model, which runs directly in the Drools engine.
It’s also possible to evaluate decision models via the Signavio’s REST APIs.
Developers of model content can publish them to a broader, non-modelling audience. Publishing a model makes it accessible in the Collaboration Hub, which provides various read-only views on the model’s content. In order to publish a model, content creators must first publish any linked Dictionary elements.
Alongside the static views of published decision models, the Collaboration Hub also features the interactive “Decision Assistant”. As shown above, The Decision Assistant provides an auto-generated form composed of all the defined input data of a decision model along with any related lists or enumerations. As the user selects or enters elements of input data, the Decision Assistant always shows the remaining set of possible results, given the partially completed set of inputs. Users can then analyse the decision details, studying the rules that fired and how the result was determined.
Import / Export
The product supports the export of decision models using the DMN XML standard. Signavio supports conformance level 2 of the standard, and some additional aspects of conformance level 3. It’s also possible to export models using Signavio’s proprietary exchange format. The import of DMN XML is under development. It’s also possible to export and import test data defined in the “test lab” element of Process Modeller in a JSON format.
Administrators set access control lists for directories, which organize and structure all modelling content. Admin users define privileges for both individual users and user groups. Integration with an Active Directory and similar means is also possible via an API.
As shown below, throughout their lifecycle decision models touch various aspects of the Signavio functionality. In terms of the direct governance of decision models – i.e. controlling signoff and changes – the Workflow Accelerator provides this sort of functionality as a part of the full Enterprise Plus version of Signavio Suite.
The approval workflow concept provides generic release management in the fully licensed Enterprise Plus version of the Suite. Customers can also define alternate approval workflows, i.e., one for each business unit.
We would like to see further support for automated governance within the tooling itself. Yes, the overall environment can control access and handle the publication of model content within the enterprise, but with powerful tools such as the Workflow Accelerator, we would expect to see it better integrated to provide automated governance of decision models and related artefacts, such as process models and terms in the Dictionary.
Of course, users can model their particular versions of common decision and process models. However, challenges emerge when one needs to start supporting variations in the Dictionary. The current Signavio approach focuses on developing a single, “standardised” taxonomy of words and meanings. In large federated organisations, this approach does not easily suit all use cases. There are times were the local context requires a localisation of the terminology. For example, when local laws apply it may not be appropriate to alter the single common dictionary item; instead, the developer might want to map the global standard to those local needs, yet still maintain the tracking. Whilst a certain degree of federated support is possible in the tool through clever use of categories and folders, it’s not particularly elegant or easy to implement as users might expect.