Topic 3 – Data / Information Architecture

POST 1 – Data Architecture: Where’s the Deviation ?

After going over a few articles this week on Data architecture best practices; what became clear to me was how much organizations deviate from best practices, despite the overwhelming amount of (freely available) information on do’s and dont’s in data.

Before I just go about regurgitating the Gatners and Forresters of the world; let me provide a simple example of a discussion we had just this week at work. Without getting too specific; here are the facts:

• System (A) is likely to be retired in the next few years
• System (B) is being positioned and developed to decouple one of many functions that was previously recorded in System A to alleviate compliance mandates.
• System A currently interfaces to System (C) for global reporting/compliance purposes
• System (B) is being built with interfaces to System (A) despite the End of Life dates of the latter staring us in the face
• System (B) seems to have left out a critical scenario to record transactions that was borne from a separate channel of interaction; thereby forcing us to continue using system (A)
• We found that the introduction of system (B) into our current application portfolio now duplicates data in 3 places just because we needed to address security and interface needs(???)

I’m sure most of the readers here can see what’s wrong with this picture. But why does this happen? Is this because of a lack of competent data architects or consultants? Or was this the result of “let’s not affect productivity in the pursuit of perfection”. Let’s just say there’s more than meets the eye here.

To go back to my point about deviation, I believe the issue with Data architecture isn’t born from bad consultants or planning. It comes from poor communication and lack of visibility. Now lets’ point to something I read in the Gartner article (you know this was coming) by David Newman about improving “Situational Awareness”. It states,

“To improve situational awareness, the enterprise’s information architecture monitors and processes diverse information from multiple sources, such as events (including unstructured content, such as news feeds), which then feed business activity monitoring (BAM) applications. Continuous streams of enterprise information enable near-real-time decision making, ensure faster response to threats or opportunities, minimize information overload, and reduce operational costs.”

How could we have improved situational awareness? What tool would have helped us improve this? Most reporting tools are based out of solutions that meet specific needs (production, logistics, sales, HR, etc…) How do we converge this information to understand application or capability overlaps/deficiencies?

In my opinion, one (of many) ways to address this would have been by leveraging an EAM tool that would easily call out the scenario I just detailed. You could have potentially seen attributes such as:

• Application dependencies (Inbound and outbound)
• Data Flows (Inbound and outbound)
• Lifecycle Management for each application
• Projects running against the portfolio
• Cost or portfolio
• Other metrics such as Business usage, Operational simplicity, Recovery times (Business Vs IT expectations), etc…

Knowing the above would have spurred effective dialogue amongst stakeholders and also addressed the communication problem. Would this have solved all our problems? Potentially, no. But we would have a better, forward looking design than what we have today.

I also read a Forrester article that speaks about how Data architecture extends to the edge of the business.

What this simply means is that data is pervasive across the organization and its functions. The recommendation is simple – maintain a steady vision on the (data) end state while supporting business outcomes.  To re-quote a quote in the article,

” Stop building projects, build programs. Momentum is the most important thing. Vendor projects were exhausting and stopped momentum. Need a continuous delivery process that shows results and validates progress”

A program to optimize our data architecture must be embedded in all our projects. This would ensure better adherence to data principles and avoid the dubious distinction of being yet another organization that deviates from Data Best practices.

References

• G, Michelle., L, Gene., E, Boris., H, Brian., M, James., & C, Elizabeth. (2017). Personalized, Fast Data Delivers Business Outcomes Your CEO Cares About. Forrester.com. Retrieved 17 September 2017
• Newman, D. (2008). Best Practices for Managing Enterprise Information. Gartner Research, (G00159499), 4-7.

 

 POST 2 & 3 – Data Strategies in Modern Age

I’ve recently been asked to be part of the Org. Change Management team in light of our recent Data Warehousing and remodeling efforts. It’s an interesting place to be if you knew about

• Historical design decisions (Where the skeletons are buried)
• The vision for how we want to use data to better help customers
• Solution impact to our Global Application and Project portfolios

I DON’T……..and here’s why I think it can (potentially) be a good thing….

Organizations with large IT teams traditionally have standards and processes defined for project/program management. The problem however, is that most processes are redundant or not as forward looking as we’d like them to be. We have often heard feedback about IT teams being too process heavy and/or lack innovative thought leadership. One of the key reason behind this is that we’re constrained by the very processes we defined to help us innovate.

When you’re a new team member to the organization, your thoughts and solution design approach aren’t as weighed down by protocol as your seasoned peers. To quote a statement from an unknown author I read years ago,

“The possibilities in the mind of a beginner are many; while those in the mind of an expert are few”

The beginner here is me – not by years of experience; but simply because of the fresh perspective our team brings to an organization that’s used to doing things a certain way. So what would be an ideal way to start when you don’t have much to go by except your own experience and a ton of project documents that would take years to read?

I decided to go back to basics. We layered that approach with knowledge available from the best research organizations out there (Gartner, Forrester, etc..) to understand critical capabilities we need in the Data Space. I particularly liked Forrester’s key criteria for an effective Data Management strategy:

1. Ability to continuously source data

   We live in a time where data is being generated an consumed at lightning speeds. We’re able to view worldwide trending topics just at the click of a button and apply this information for strategic business advantage. Capability to view live insights for sales, marketing, service data is critical in today’s world when brand loyalty is a thing of the past.

2. Capture Data with the right context

   Everyone has data. Tons of it. The problem is we don’t know what to do with it & we’re sitting on a treasure trove of information without knowledge of how to effectively draw insights from it. I’ve personally seen many organizations just stash away or ignore their data when migrating from one platform to another because it was too expensive to migrate it over. Services like AWS or Azure today provide cost effective means to store historical data while you figure out your data strategies.

3. Data enrichment through scientific approach

   Using technologies that help you weave context to your data or automatically provide insights based on your information to revitalize or improve your products/services is necessary for organizations looking to differentiate themselves from competitors. Enriching your customer profiles using Social Media and intelligent analytics help you better position your products or improve them.

4. High speed data delivery on demand

   My favorite. We are living in a time where reporting is no longer a long, time-consuming process. We have technologies that help us spin up reports and dashboards in minutes using drag and drop tools. Cloud based solutions are no longer bound by limitations of their on-premise, hardware bound peers. Management teams should be able to draw new insights from their existing data on demand to respond to competitive pressures. We often hear news about how sales of product in a certain category go up or down in response to a planned or unplanned global/national events. An organization that’s ready to take advantage of such events spin up effective marketing, promotions and even ramp up production schedules to avoid missing sales opportunities OR even just use the opportunity to improve brand awareness. e.g.  Tesla temporarily unlocking full battery capacity for those affected by hurricanes recently makes this an excellent case for customer’s looking to pay the money for a permanent unlock and increase their brand satisfaction because Tesla was by their side when they needed it the most.

The above criteria is exceptionally insightful for organizations looking to revitalize their Data Management strategies. I’m sure many organization will have to make their own unique tweaks to make this work for them; but the end results are a strategic view to help address data in the modern age.

POST 3 Continued – The HOW

Forrester have laid out 4 simple steps you can take to make this a viable journey:

1. Assemble the right stakeholders
   This involves gathering critical decision makers and strategic advisers who can provide real, unbiased insight into what our current data capabilities are and why. We can use these individuals/roles to understand the pitfalls of past design decisions and use any available tribal knowledge to avoid mistakes of the past. Recommended stakeholders are Chief Data Officer, Data Governance Program leaders, Enterprise architects, Business Data Stewards, Data Governance leaders, Business Analysts. Many organizations dont have formal positions or roles listed here; but that can be used to trigger change or extend the role of existing resources.
2. Perform a needs assessment
   Needs assessments are a topic that has been discussed so often that the words themselves triggers groans and sighs. I’m not saying the words need to change; but the strategic message around needs assessment and the positive impact each role brings to the table needs to be amplified. Needs assessment is a tool, not a job.
3. Investigate your strategic options
   Forrester talks about a few interesting points here, but I ‘ll pick the one that’s most important in my opinion. The most insightful point is to be selective about who is on your strategy team. I’m aware it sounds bland and obvious; but there’s more to it. In Forrester’s view the team should be not more than six members; half of which belongs to business as opposed to IT. These business members should be those who can represent sales, marketing and other functions equally. It also speaks about how teams can augment themselves with additional external support for skills capabilities they lack while owning their strategic function. If implemented well, this reduces a lot of cross shopping of solutions in larger organizations. Yet another suggestion is to oscillate between viewpoints of top-down and bottom-up to see if strategic needs and processes put in place are being pursued by operational processes (diagram below).
   
4. Prioritize your strategic options
   The last step in the process helps all stakeholders identify the strategic options they want to pursue in light of the capabilities that they seek. The exercise is intended to help all involved take a pragmatic view at all the options that can be re-used in its current state, integrated with state of the art technologies available and finally decide on an approach that will set the road-map for the next 10 years.

The outcome here is twofold:

• You learn what to do with those skeletons I spoke about earlier
• You recognize that being a beginner sometimes isn’t such a bad thing after all

References

• G, Michelle. (2017).Four Steps to Data Management in Light of Big Data. Forrester.com. Retrieved 24 September 2017
• TOGAF 9.1 (2008). Data Architecture Principles.