IBM Cloud Strategy: Collaboration, Dev/Test Environments, and Virtual Desktops

Today, IBM announced their cloud strategy and roadmap; I was at the analyst update last week and had a chance to hear about it first-hand from IBM execs, a customer and a partner.

Erich Clementi, who heads enterprise initiatives at IBM, started the briefing by showing their cloud evolution over the past year, and plans for the remainder of 2010. Last year saw the launch of LotusLive collaboration services and the Test Cloud for hosted test environments. By the end of 2009, cloud offerings had expanded to include analytics, storage and email plus cloud consulting services, and the beta for cloud-based development and test environments had opened up. That beta has evolved so that today we’re hearing about the launch of Smart Business Dev/Test on IBM Cloud: an enterprise-class environment for provisioning virtual machines on demand for software development and testing. By the end of this year, there will be more cloud offerings, and a variety of security, resiliency, capacity and compliance options, and an ecosystem of partners.

He discussed what they’ve learned from their clients: there is a universal interest in cloud computing, but that there won’t be a “Big Switch fantasy” happening in large enterprises any time soon. Instead, this is part of a transition from owning IT assets to sourcing IT solutions as part of an organization’s enterprise IT delivery mix, where cloud complements on-premise, and these often coexist in integrated hybrid services. Although cost is a factor, speed of deployment is also a key driver, since that drives time to value. And, since IBM always has a large services component, they have a suite of services around moving to and maintaining cloud services. To be clear, there is a predominant focus on private clouds, or what some would not consider cloud at all: fast provisioning (after you install all the hardware and infrastructure software), but everything is on the customer’s site, making this virtualization rather than true hosted cloud.

For hosted cloud, they see the initial sweet spot as the collaboration space, where they’re targeting the LotusLive brand, including the web conferencing tool which we were using for the briefing, email suite (Lotus Notes lives!) and even social networking, such as the BPM BlueWorks community. Altogether, IBM has 18 million users on LotusLive, including their own workforce and some large customers such as Panasonic.

Targeting both public and private cloud is their Smart Business Desktop, where the entire desktop environment – OS and applications – is virtualized rather than installed on the actual desktops, allowing for access from anywhere, and also providing desktop remote control and other IT service functions. This has long been used for VPN access to networks, but is a newer concept for full-time internal desktops. Coincidentally, eWeek just published an article on virtual desktop infrastructure (VDI), discussing the benefits in terms of reduced maintenance and hardware costs (reduce desktop TCO by 15-35%) as well as business continuity, but also the relatively high startup costs and complexity; the author ultimately states “I hesitate to recommend VDI across the board”.

The third part of their cloud strategy is for virtual hosted server environments for ISVs – what appears to be a direct competitor to Amazon’s EC2 – providing development and test infrastructure through developerWorks Cloud Computing Resources, but apparently also production hosting (I think – the presentation was a bit vague here).

For my regular BPM readers, if you’ve made it this far, consider how you could use cloud development and test servers for BPM deployments, where some of the multiple environments required (usually at least four, sometimes as many as six) could be moved out of your own data centers, and provisioned at will.

Pat Toole, CIO of IBM, was up next to discuss how they are using their own products internally, speaking as a customer of the cloud offerings. They started with hosted development and test environments, and have half of their new development in the US happening on the dev/test cloud; this has reduced their server provisioning time from five days down to about an hour for both Power and x86 environments. Next, they looked at BI and analytics, with the dual aim of reducing costs and making the data more readily available to users. They consolidated 100 data warehouses into a single Cognos environment for 80,000 internal users in their Blue Insight initiative, and expect to add another 30 applications and double their users over the next year.

On the collaboration front, they turned on LotusLive web conferencing for all employees to use for internal and external meetings, logging 200 million minutes last year. They’ve recently added Engage for 6,000 users initially; although this seems to provide full social networking capabilities, Toole mentioned file sharing as the primary use case.

They’ve implemented Smart Business Desktop at one center in China in order to reduce TCO by more than 40% and improve security and control, and plan to roll this out to their call centers in US and India. Echoing the eWeek article, he said that this is not for everyone in the organization, but makes sense for certain classes of users and desktops. They’re also about to launch their first pilot on the storage cloud, and have identified about 1,000 applications for deployment in the production cloud.

In “eating their own cooking”, IBM is doing what any of their customers would be doing: trying to make their computing environment more efficient and less expensive.

Mike McCarthy, who heads the cloud computing group, gave the the details of today’s announcement:

  • Smart Business Development and Test environments on the IBM (public) cloud, initially within North America, on a pay-as-you-go or reserved capacity basis. Although hosted on their public cloud, this is intended to support enterprise clients in that it’s not an open community, but a platform for hosting your development and test environments as securely as if they were on premise; in fact, they plan to offer dedicated hardware environments in the future for the truly paranoid. There are several pre-configured software images to select from, offering a wide choice of configurations and deployment models. They offer 99.5% availability, sufficient for most dev/test environments, and support options up to 24×7 telephone support. This allows you to provision a development or test environment yourself in a matter of minutes: choose the service (software image, such as OS or OS plus tools), configure the usage configuration, and click to provision a new virtual server. Initially, they’ll be offering Red Hat and Novell Linux on x86 environments, with additional hardware options as well as Windows later in the year.
  • Adding development services, such as Rational SDS, to the existing Smart Business Test Cloud offering for private cloud deployments.
  • Rational Software Delivery Services for both their private and public Smart Business Development and Test Cloud.
  • Tighter integration of the developerWorks online community and the development/test cloud initiatives through a variety of learning resources.

Evan Bauer of the Collaborative Software Initiative joined the IBM team on the call to discuss their use of the IBM cloud for developing, testing and hosting the US Department of Education’s Open Innovation Portal. They used the beta version of the IBM cloud and open source software to develop and deploy this portal within three months. Hosting on IBM’s public cloud allows them to scale quickly and achieve excellent response time, providing a valuable pilot for the future use of cloud for government applications.

Last up was Tom Lounibos of SOASTA, an IBM partner offering CloudTest, an on-demand service for load-testing web applications by provisioning hundreds of virtual servers to simulate millions of users hitting a website. There are a couple of key use cases for this type of load-testing – e-commerce sites with seasonal peaks, and social media sites with peaks caused by news events – with some very high profile cases of unacceptable latency or even site failure due to load. CloudTest has been around for a while, but they’ve just announced that they’ll be running on the IBM cloud.

The IBM (public) cloud will initially be hosted in the US, with data centers in Europe added later in 2010. Although there was some talk about other data centers (such as Asia) in the future, the entire rollout plan wasn’t clear. Many organizations, especially financial services, need to have the data centers located in their own country, or at least one with better privacy laws than the US, so both the location of the data centers and the ability for a customer to select which country is hosting their systems will become important as IBM looks beyond the US market.

For those of us used to working with virtual servers hosted elsewhere, the concepts announced today aren’t new, but the IBM brand brings an air of respectability to the idea of using hosted virtual environments for a variety of uses.

ARISalign Online Process Modeling and BPM Community

There has been much speculation in the BPM world about Software AG’s online BPM community, originally dubbed AlignSpace, or as it has been recently renamed, ARISalign. Originally launched in a private beta months ago, those of us on the outside have been anticipating a look at how they plan to “combine social networking tools with intuitive tools for process design and modeling [to] collaborate effectively to create and improve processes”.

A few weeks ago, prior to the official beta release, I had a chance for a briefing with Thomas Stoesser of Software AG for a closer look, and I’ve been playing around with it myself since the beta opened. With ARISalign, they’re providing tools for collaboratively capturing business processes in an early process discovery stage, and also providing an open BPM community for anyone to participate, not just ARIS and webMethods users. In the future, they’re also planning for a marketplace for BPM-related products and services, although that’s not in the current offering.

Home screenLogging in to ARISalign, you see a home dashboard that shows a feed of updates on your projects, groups, discussions and networks, plus a message center and a list of your current projects. There’s also a Facebook-like status feature, although I’m not sure that I’d use this feature since it’s unlikely to be my primary social network – I don’t even do Facebook status updates any more since I started Twittering.

Projects are how process artifacts are organized in ARISalign, with a project including a number of components:

  • A whiteboard, similar in appearance to Lombardi Blueprint and other process discovery tools, that allows users to add “stages”, then activities that belong to each stage. 
  • Any number of process maps that can be linked to, but not generated from, the activities on the whiteboard.
  • A discussion forum, which provides a simple threaded discussion board within the project.
  • A library of related files/documents that can be uploaded as background or reference materials. Currently, the library can only contain uploaded content, not links to content that is hosted elsewhere; links would have to be added in a discussion thread.

If you like the project framework but don’t plan to add process models, then a group has all the same features as a project except for the whiteboard and process maps: you can use it if you want only a discussion forum, library and timeline shared between a group of people.

Creating a project requires only specifying a project name: everything else is optional or has some reasonable defaults. You can add a description, and select industry and language from predefined lists, although these are used as project search metadata only and don’t change the form of the project in any way. You can also select the access control for viewing the project, confusingly called “Project Type”, as open (visible to all), restricted (anyone can see the project in a search list, but not the details or content) or hidden (not visible to non-members, even in search results). All projects require that you join the project in order to participate, which may or may not require a process administrator’s approval.

There are three roles that a member can be assigned for a specific project:

  • Project administrator, including the project owner/creator, which allows all functions including administering members, changing user roles, and archiving and renaming content.
  • Project contributor, which allows working with tools and adding content.
  • Project reviewer, which allows viewing content, participating in discussions and adding comments, but not changing content such as process models.

Unfortunately, there is no way to change the project owner from the original creator, although this is in the future plans, as is the idea of creating project templates for faster startup.

For an existing project, members will often want to start on the project dashboard where they can view a feed of all activity on the project (echoing the personal dashboard for a user, which shows activities for a user, their projects and their network). Similar to functionality recently added to Facebook, a user can hide specific people, models and activities on the dashboard, which creates a filter of only their view, not everyone’s view of the project dashboard.

Comments indicator on activityTo get started with process modeling, however, you’ll start on the project’s whiteboard tab, a near-real-time collaborative process discovery tool. High-level process steps, or stages, are added, then activities added below each stage: a process discovery paradigm for non-process-oriented users to just list the steps that are involved in the process. All project members can see each other’s changes as they occur, and can invite additional project members directly from the whiteboard view. Activities can be assigned properties, including comments by project reviewers; activities with comments show a pencil icon on the activity so that others know that comments exist. In the future, activities will also be able to have attachments; currently, attachments can only be added to the project library.

The whiteboard view also allows adding goals and KPIs, although these are purely informational and can’t (yet) be applied to any process models created within that project. In the future, there may be value in considering how KPIs can be linked to the process models and exported for use in other tools.

Unlike some other process discovery tools, the whiteboard view does not auto-generate a process model – apparently there was quite a bit of internal design conversation over whether to do this or not – but one or more process models can be added to the project. Adding and editing a process model creates a split screen view with the whiteboard and the process model; activities can be dragged from the whiteboard to the process model, which creates a linkage between the activity in those two locations, such that highlighting the activity on the whiteboard also highlights it on the process model, and vice versa. Swimlanes and subprocesses in proces view - also, selecting linked activity in either view highlights bothA whiteboard activity may be linked to more than one process model, so changes to the activity are not promoted to the process model. There can also be whiteboard activities that don’t end up on any process model. I’m not sure that I’m on board with this method; first of all, I would like to see a way to auto-generate a process model from the whiteboard, and I also think that if something is in the whiteboard view, it needs to be on a process model somewhere: otherwise, why is it in the whiteboard view at all? It appears that the reasoning behind this is that the process model is intended to be an executable process model, such that only the things that might end up in a BPMS would be included, whereas the whiteboard model includes purely manual tasks. Multiple processes from one whiteboard appears to make sense so that non-process people don’t have to think about what are distinct processes, but on second glance, I’m not sure that’s the right way to go.

The more we dig into this, the more that I’m left with the feeling that this is a front-end for webMethods, not an ARIS extension, although the process modeling palette looks more like ARIS Express rather than the webMethods Designer. ARISalign is intended to be a purely business tool, so doesn’t expose web services calls or other technical complexities.

Process models can be exported to webMethods format, XPDL, or opened directly in ARIS Express, but there’s no round-tripping since importing the model back from ARIS Express requires uploading it as a different project. ARIS Express now supports “whiteboard” collaborative models, so the whiteboard can be exported and opened in ARIS Express as well as the process model. There are no offline capabilities; the only offline alternative would be to export to ARIS Express, then upload the changed models to a different project or take screen snaps of the ARIS Express changes and add as images to the project library to document offline changes. Neither of these is particularly attractive, so this may not be an option if you have to have offline access. There are plans to improve the ARIS Express integration in the future, possibly allowing a process model to be downloaded and locked for editing in ARIS Express, then re-uploaded in place.

There’s a view of all process models in a project, which allows those models to be managed (renamed, exported, deleted), but any editing of the models occurs in the split-screen view with the project whiteboard.

Recommendations for connectionsAside from the project functionality, there are a number of social networking features for managing your profile and your connections. You can set different views of your profile for your network or for public display, and can view recommendations of people to whom you might want to connect based on company, industry and shared contacts. The Message Center is very Google Wave-like, with participants shown at the top, and allowing public or private reply to any part of the thread. This holds potential to become the conversation framework used within projects, to replace the current simple discussion groups. In general, the UI is quite nice (although some may not like that it was created with Adobe Flex), and has borrowed liberally from successful features of Facebook and other social networks. The navigation is quite flat, making it easy to find your way around in the interface.

Software AG also showed off an ARISalign iPhone app at CeBIT, although it’s not generally available yet. I’m not sure I’d use this for much process modeling, although it would be useful for tracking what’s happening on projects, accepting invitations, participating in discussions and even looking at (or some light editing of) the whiteboard view.

Currently, ARISalign is available only as a hosted solution, and is hosted on the US version of Amazon Web Services. It’s architected so that on-premise hosting could be enabled in the future, although not in the current version. Software AG should consider having a version hosted on the EU AWS instance, since many organizations don’t want their information – even process models that don’t contain customer data – hosted in the US due to the privacy laws.

This is the first publicly-available version of ARISalign, and no one expects it to be perfect. How quickly Software AG can respond to users’ requests for new functionality – such as the inclusion of a marketplace for add-on applications and services – will be the real test of success, as I mentioned in my recent review of the IBM BlueWorks community.

There’s also the issue of merging the existing ARIS Community with ARISalign or at least cross-linking user accounts, which seems a logical step, but is not permitted by Germany privacy laws until Software AG and IDS Scheer officially become a single company, which could be several months still. The two sites may not end up merged; you can imagine the ARIS Community site being left with product support for ARIS and remain more actively managed, while the user-generated content such as discussions as well as the more generic tools be moved over to ARISalign. You can be sure that there will be some internal politics around this decision, too. Regardless, in the mean time, there’s a badge in the sidebar of each site linking to the other, encouraging you to sign up on the other site. That might, however, cause a bit of social networking fatigue for many business users.

ARISalign

Paper on Runtime Collaboration and Dynamic Modeling in BPM

I recently wrote a paper for the February Cutter IT Journal called Runtime Collaboration and Dynamic Modeling in BPM: Allowing the Business to Shape Its Own Processes on the Fly. It’s available on the web to Cutter subscribers, and in the printed journal.

In the article, I deal purely with the topic of runtime collaboration, not collaboration during process modeling: how users participating in a process can add new participants for the purposes of collaborating on a step in a structured process, or even create their own subprocess at that step. I look at why you would want to do that (mostly auditability of processes) and how the results of that can be rolled back into process design rather than just being changes to a single process instance.

Disclosure: my payment for writing this paper was a year’s subscription to the journal, plus bragging rights.

IBM BlueWorks Online BPM Community

I had a briefing a couple of weeks ago on IBM BlueWorks by Angel Diaz and Janine Sneed from the BlueWorks team. BlueWorks is IBM’s cloud-based BPM environment, providing the following capabilities:

  • Browser-based modeling, including strategy maps, capability maps, process maps and BPMN processes.
  • Pre-built content to supplement or replace a BPM center of excellence (CoE), including the ability to submit your own content.
  • Online community for collaboration and exchange of ideas.

BlueWorks content viewBlueWorks was launched last July, and has several thousand people signed up, although I didn’t get a good feel for the level of activity. It’s based on Lotus Business Space, with the modeling editor and repository from the WebSphere BPM suite, which allows IBM to offer both a hosted and on-premise version.

They’ve kicked off the content part of BlueWorks by seeding it with a lot of content from internal and external contributors, including information provided by their professional services arm. The results is a large repository of articles, sample strategy maps, business measures such as KPIs, forums and blogs with more information that you could hope to scavenge through. It’s all categorized and tagged in multiple ways, however, making it easy to filter the library to just what you’re looking for, whether by topic, industry, or type of content. They also include industry content packs, which are bundles of industry-specific strategy maps and other content.

BlueWorks process modelThe process designer is Flash-based, and it only took me about 5 minutes to crash it; luckily, it saved as I worked, so I didn’t lose any work. Some of the operations are not very intuitive (I had to go to the help file to figure out how to add a new activity), but once I learned a few of the basics, it’s pretty efficient to use, and I could use the keyboard for entering my activity list, which I like. The process is shown in both a text outline view and a process outline view, very similar to other process discovery/outlining tools such as Lombardi Blueprint (which should make the integration of Blueprint into this environment straightforward from a user interface standpoint, if not a technical one). Once complete, I could export to a PowerPoint presentation (which includes slides for the process model and the details that I entered), a Business Document Archive (a binary format that I’m not familiar with) to my local file system or the asset repository, or to a WebSphere Business Modeler XML format.

BlueWorks BPMN modelThis is where I found things a bit strange: I couldn’t export or otherwise convert the process model that I had created to use in the BPMN modeler, which is a separate tool. Maybe this is something that the Blueprint folks can teach them about. I found the BPMN modeler a bit clunky: resizing and placement of elements was awkward, although it allowed me to validate my model as valid BPMN. There definitely needs to be a way to move between these two process model types, to eliminate redrawing and also to allow a process analyst to quickly flip between the different perspectives. From the BPMN model, I could save to the shared repository, or export to BPMN 2.0 XML, WebSphere Business Modeler XML, or a Process Diagram Archive XML format.

I didn’t spend a lot of time on the strategy or capability maps; a strategy map is a mind-map type of model that allows you to model business SWOT factors as well as business goals, whereas a capability map shows the business capabilities and can link them to process models. The strategy, capability and process maps all have a similar user experience, and are all shown as siblings within folders in the BlueWorks space under the Design tab; BPMN models, on the other hand, are shown in a separate tab and have a completely different UI. The BPMN model seems like a bit of an add-on: obviously, there’s a need for BPMN modeling in an online BPM community, but they haven’t quite got it integrated yet. The three Design map types are really intended for business users, and allow functions such as pasting an indented bulleted list from a PowerPoint presentation into a strategy map to create an initial map. Links and attachments (including documents and folders) can be added to any node in any of the three Design diagram types. All four model types have versioning, and models of all types are visible in my dashboard view.

BlueWorks share model dialogAside from the functionality of the modelers, there’s the ability to collaborate on models: each person has their own private space in BlueWorks, or they can share their models with their team members. The upcoming version 7 of BlueWorks will allow more fine-grained privacy controls to allow sharing only with specific groups.

The content and community parts of BlueWorks form the basis of a CoE: smaller companies could use this as their only CoE, whereas larger ones might want to use content from BlueWorks with their own internal content. Content submitted to the content section is not only visible to anyone on BlueWorks, but also is explicitly licensed to IBM for redistribution, so this isn’t a place for your private intellectual property, but a good place to share ideas with people from other companies. IBM partner companies are starting to use it for sales material and starter content.

The hosted version of BlueWorks is free, and you don’t even need to be an IBM customer, but if you want to take this capability inside your own firewall, IBM would be happy to sell you WebSphere Business Compass (formerly WebSphere Publisher). Also based on Business Space, Diaz described it as an in-house version of BlueWorks, but it has many more tools such as forms designers, organization charts and other process modeling tools. You don’t need to use WebSphere Business Compass – it’s possible to go directly from BlueWorks to an executable system using the WebSphere and BPMN export formats – but for some companies, BlueWorks will act as the “gateway drug” to get them hooked on the bigger and better functionality of Business Compass.

I was briefed on Software AG’s online community, ARISalign, earlier this week and will post my thoughts on that soon; in both cases, these competing online communities lack some key functionality, but need to get their platforms out there for people to start using and feeding back on what’s needed. The best online community will result not from who has the most advanced starting point, but from who can be most responsive to their community’s needs.

You can sign up for your own BlueWorks account for free, and there’s a webinar tomorrow at 1pm ET on getting started with BlueWorks that will be recorded and available for replay later.

Enterprise 2.0 Adoption Council

In my conversations with several people this week, I’ve mentioned the Enterprise 2.0 Adoption Council as a place to start for finding out information about companies that are actually implementing this stuff. The Enterprise 2.0 Adoption Council gathers together managers from large organizations who are spearheading these efforts inside their company. The site includes research that the Council is doing on its members, such as their “State of Enterprise 2.0 Adoption” report for Q4 2009. Susan Scrupski conceived the idea for the Council and pulled it all together. You can follow her on Twitter here and read her business blog here for lots of additional information on how Enterprise 2.0 is being used in large enterprises.

If you have an interest in this area, then lots of good information here. If you’re a large enterprise implementing – or even looking at – Enterprise 2.0, consider joining the Council.

BPM and Enterprise 2.0 at Software 2010 in Oslo

I’m at the Software 2010 conference held by the Norwegian Computing Society in Oslo this week, and gave the opening keynote on one of the tracks this morning: how Business Process Management is being impacted by social software and social networking:

I gave a similar talk last November at the Business Rules Forum, but I find this topic to be endlessly changing and endlessly fascinating. I’ve written two related papers on it recently, too: one for the Springer BPM Handbook, and one for the Cutter IT Journal (specifically on runtime collaboration in BPM).

I won’t be attending most of the other sessions because they’re in Norwegian, but may pop out this afternoon and visit the Edvard Munch works at the National Gallery. I spent a few days in London earlier this week, visiting the Victoria & Albert, Tate Modern and British Museums, so that would round out my week nicely.

BPM, Collaboration and Social Networking

Although social software and BPM is an underlying theme in a lot of the presentations that I give, today at the Business Rules Forum is the first time that I’ve been able to focus exclusively on that topic in a presentation for more than 3 years. Here’s the slides, and a list of the references that I used:

References:

There are many other references in this field; feel free to add your favorites in the comments section.

Just call me “Your Honor”

Apparently, Shel Israel’s fact checkers were too busy to actually check facts the day that they proofed page 208 of his new book Twitterville: I am not, nor have I ever been, the mayor of Toronto.

After a couple of people alerted me (via Twitter, of course), I hiked over to the local bookstore and snapped a pic of the page in question – click to see the full-size image, and check under the heading “Tweeting International” near the bottom left where it refers to “Toronto mayor Sandy Kemsley (@skemsley)”. I didn’t buy the book: if it lists me as the mayor of Toronto, who knows what other nonsense it contains? 🙂

fyi, the mayor of Toronto is David Miller (at least until the next election), a.k.a. @mayormiller. I am, however, one of the 57 people who he follows.

Getting Business Process Value From Social Networks #GartnerBPM

For the last session of the day, I attended Carol Rozwell’s presentation on social network analysis and the impact of understanding network processes. I’ll be doing a presentation at Business Rules Forum next month on social networking and BPM, so this is especially interesting even though I’ll be covering a lot of other information besides social graphs.

She started with the (by now, I hope obvious) statement that what you don’t know about your social network can, in fact, hurt you: there are a lot of stories around about how companies have and have not made good use of their social network, and the consequences of those activities.

She posited that while business process analysis tells us about the sequence of steps, what can be eliminated and where automation can help, social network analysis tells us about the intricacies of working relationships, the complexity and variability of roles, the critical people and untapped resources, and operational effectiveness. Many of us are working very differently than we were several years ago, but this isn’t just about “digital natives” entering the workforce, it’s about the changing work environment and resources available to all of us. We’re all more connected (although many Blackberry slaves don’t necessarily see this as an advantage), more visual in terms of graphical representations and multimedia, more interactively involved in content creation, and we do more multitasking in an increasingly dynamic environment. The line between work and personal life blurs, and although some people decry this, I like it: I can go to many places in the world, meet up with someone who I met through business, and enjoy some leisure time together. I have business contacts on Facebook in additional to personal friends, and I know that many business contacts read my personal blog (especially the recent foodie posts) as well as my business blog. I don’t really have a lot to hide, so don’t have problem with that level of transparency; I’m also not afraid to turn off my phone and stop checking my email if I want to get away from it all.

Your employees are already using social media, whether you allow it within your firewall or not, so you might as well suck it up and educate them on what they can and can’t say about your company on Twitter. If you’re on the employee side, then you need to embrace the fact that you’re connected, and stop publishing those embarrassing photos of yourself on Facebook even if you’re not directly connected to your boss.

She showed a chart of social networks, with the horizontal axis ranging from emergent to engineered, and the vertical axis from interest-driven to purpose-driven. I think that she’s missing a few things here: for example, open source communities are emergent and purpose-driven, that is, at the top left of the graph, although all of her examples range roughly along the diagonal from bottom left to top right.

There are a lot of reasons for analyzing social networks, such as predicting trends and identifying new potential sources of resources, and a few different techniques for doing this:

  • Organizational network analysis (ONA), which examines the connections amongst people in groups
  • Value network analysis (VNA), which examines the relationships used to create economic value
  • Influence analysis, a type of cluster analysis that pinpoints people, associations and trends

Rozwell showed an interesting example of a company’s organizational chart, then the same players represented in an ONA. Although it’s not clear exactly what the social network is based on – presumably some sort of interpersonal interaction – it highlights issues within the company in that some people have no direct relation to their direct reports, and one person who was low in the organizational chart was a key linkage between different departments and people.

She showed an example of VNA, where the linkages between a retailer, distributor, manufacturer and contract manufacturer where shown: orders, movements of goods, and payments. This allows the exchanges of value, whether tangible or intangible, to be highlighted and analyzed.

Her influence analysis example discussed the people who monitor social media – either within a company or their PR agency – to analyze the contributors, determine which are relevant and credible, and use that to drive engagement with the social media contributors. I get a few emails per day from people who start with “I read your blog and think that you should talk to my customer about their new BPM widget”, so I know that there are a lot of these around.

There are some basic features that you look for when doing network analysis: central connectors (those people in the middle of a cluster), peripheral players (connected to only one or two others), and brokers (people who form the connection between two clusters).

There are some pretty significant differences between ONA, VNA and business process analysis, although there are some clear linkages: VNA could have a direct impact on understanding the business process flows, while ONA could help to inform the roles and responsibilities. She discussed a case study of a company that did a business process analysis and an ONA, and used the ONA on the redesigned process in order to redesign roles to reduce variability, identify roles most impacted by automation, and expose critical vendor relationships.

Determining how to measure a social network can be a challenge: one telecom company used records of voice calls, SMS and other person-to-person communications in order to develop marketing campaigns and pricing strategies. That sounds like a complete invasion of privacy to me, but we’ve come to expect that from our telecom providers.

The example of using social networks to find potential resources is something that a lot of large professional services firms are testing out: she showed an example that looked vaguely familiar where employees indicated their expertise and interests, and other employees could look for others with specific sets of skills. I know that IBM does some of this with their internal Beehive system, and I saw a presentation on this at the last Enterprise 2.0 conference.

There are also a lot of examples of how companies use social networks to engage their customers, and a “community manager” position has been created at many organizations to help manage those relationships. There are a lot of ways to do this poorly – such as blasting advertising to your community – but plenty of ways to make it work for you. Once things get rolling in such a public social network, the same sort of social network analysis techniques can be applied in order to find the key people in your social network, even if they don’t work for you, and even if they primarily take an observer role.

Tons of interesting stuff here, and I have a lot of ideas of how this impacts BPM – but you’ll have to come to Business Rules Forum to hear about that.

AGILIPO: Embedding social software features into business process tools #BPM2009 #BPMS2’09

Three years ago, I gave a presentation entitled “Web 2.0 and BPM” at the BPMG conference in London, in which I said that the future of BPM and Web 2.0 (or what we would now call social software) will include tagging of process instances. Today, I saw some research that includes exactly that functionality, as part of a presentation on AGILIPO, an environment based on Agile principles to provide collaborative process modeling and execution. I missed the presenter’s name, and with six authors on the associated paper, I’m not going to take a guess. Update: the presenter was David Martinho of the Center for Organizational Design and Engineering in Lisbon.

This was a very fast-paced presentation with an Ignite-like pacing, so it was difficult to take notes, but there was some great material here. It really is looking at an Agile approach, with “short feedback cycles to steer instead of align” to avoid long modeling phases and allow unexpected behavior to occur.

One interesting bit is on allowing a user to create generic (non-specific) activities: what we’re starting to see in the commercial market as dynamic BPM, but simplified by allowing the inclusion of generic activities that can then be tagged with their specific details. The user adding the activity will assign tags specific to this process instance to associate semantics to the dynamically-added generic activity; the new process steps can be added back to the main process design through versioning.

Both comments and tags are allowed at the process instance level, as well as on the step in the process model.

Interesting ideas; I look forward to seeing how the concepts of process instance tagging are used in other applications.