Warning: "continue" targeting switch is equivalent to "break". Did you mean to use "continue 2"? in /homepages/8/d158560430/htdocs/ericjlawson/wp-content/plugins/jetpack/_inc/lib/class.media-summary.php on line 77

Warning: "continue" targeting switch is equivalent to "break". Did you mean to use "continue 2"? in /homepages/8/d158560430/htdocs/ericjlawson/wp-content/plugins/jetpack/_inc/lib/class.media-summary.php on line 87
Super(hung out to)Dry | Data Architecture
Warning: count(): Parameter must be an array or an object that implements Countable in /homepages/8/d158560430/htdocs/ericjlawson/wp-includes/post-template.php on line 284

Super(hung out to)Dry

The admission in April 2012 by SuperGroup (the parent company of fashion brand SuperDry), that the entry of a ‘+’ sign instead of a ‘-‘ sign on some forecasting figures, led to its 3rd ‘profit’ warning being issued in 6 months. This led to a revised annual profit shortfall of £8m. Robust business intelligence (BI) processes might have prevented this data error.

At the time a city analyst declared the group’s finance department ‘not fit for purpose’. Its shares tumbled 38% on this news. A month ago, the group announced the recruitment of a new COO (new role), a new CFO and a new Head of Retail in an effort to boost the management team and increase the professionalism of the group.

In general terms IT provisioning for a business such as SuperGroup falls into two broad categories. The first are the type of solutions that support the procurement and sale of goods, wholesale or retail. This is traditional IT and it has been around for a long time, delivering improvements and cost savings in areas like HR, Supply Chain, Marketing etc. New sales channels and models have developed over time (eg. the web, eBay etc), but the focus is still on response times, availability and scalability. It’s all about getting the job done, modeling business processes and providing automation and capture systems as appropriate.

The second category of IT provisioning is centred on decision making. Business relies very heavily on information but not all are fortunate enough to have timely and reliable data to work with. Manual number crunching and intuition (“gut feel”) have served many a business well but it’s not enough anymore, especially as a business grows beyond the point where everybody knows each other by first name.

Traditionally, information for decision making in larger businesses was delivered either via the core ‘online’ system or through nightly or cyclic report production runs. Information delivered via online systems generally took the form of single record data items, such as information about the customer, supplier or order that a clerk / call centre operator might be queried about. Reports were run to arrive at the summary totals on the last page, to assess the outcome of the previous day’s business, for order picking, invoicing etc.

Over the last 10-15 years operational data stores and data marts came into vogue with the purpose of feeding derived business intelligence back into the core online systems. Combined with the transition from dumb terminals to PC based client interfaces, IT was able to arm the customer facing areas of the business (and other areas too) with as much relevant information about the “customer” as possible.

Decision support systems began to get airplay in IT departments. Data Services teams were formed to handle the growing number of multi purpose databases and data oriented solutions. The methodologies of data warehousing advocates Ralph Kimball and Bill Inmon, gained significant traction. Programs specific for moving data between disparate systems and databases, got the new moniker Extract, Transform & Load (ETL).

However, still the primary focus of many IT departments was on the core online systems, with a centralised infrastructure and traditional project methodologies. And fair enough too as there has been plenty to contend with. This would be especially true for small, rapidly expanding businesses. However, I am sure there a many exceptions to this generalisation of focus and perhaps some of them are now Fortune 500 companies (and the equivalent across the world markets).

Solutions deployed were quite often ‘brittle’ (a term I heard Matt Benati of Attunity use recently and I think it’s quite apt). That is, not quick to adapt to changes in the business and prone to breaking. An architect I have had the fortune of working with lately has been advocating significant levels of isolation to shield downstream systems (such as a data warehouse) from the impact of change. However, it’s not always easy or practical to do this.

Another new term has been cropping up recently (Attunity again), in the big data discourse. Extract, Expedite and Load (EEL). It’s interesting because the premise here is on getting data out of capture systems in real time and consuming it as fast as possible in the target system. This concept is very compelling. I personally think that it will gain momentum.

The SuperGroup example is a timely reminder that robust BI processes are critical to any organisation operating in todays business environment. A rapidly growing business should treat BI with equal or more importance than core IT systems. Aggressive business expansion without commensurate capability to measure those business activities themselves, is a recipe for disaster. “Reporting” as BI is often lazily categorised, should not be 2nd string. Failure may not be as immediate and obvious as an “offline” order taking system, but it can be just as damaging to the bottom line (or more so).

Larger organisations needs to get tough with in place IT power bases. Core IT systems require stability combined with continuous improvement. BI solutions are a completely different proposition, requiring many traditional IT skills & features. When combined with some of the new amazing new database solutions and storage advances (column based indexing, SSD’s, cloud deployments, CDC software etc), significant improvements can be made.

I am not talking about traditional operational reporting (which can be a deliverable of the BI team). That’s the straightforward, routine, but bottom line piece of the puzzle. BI solutions must also take the tactical and operational goals of the business (derivable from the strategic vision and plans) and provide an integrated view, all the way from executive dashboards, through tactical performance analysis down to operational workbenches and shop floor devices (and all the way back to the top).

However, the often overlooked challenges for BI are the human ones, the process based issues. This links back to how IT systems development needs to embrace and inform business process. Unfortunately this is when resistance to change is encountered, throughout all levels of the business. This was a factor in the 80s and 90s and clearly remains relevant today. Back then the change mainly impacted the front line and more blue collar oriented workers. Now, transitional BI impacts the strategic & tactical decision makers.

There needs to be acknowledgement of the direct and crucial relationship between steering a business through a rapidly changing landscape and the continuous need for timely and accurate data to support the process. If a business is forced to make rapid changes, the supporting data needs to be delivered into the BI system now, sometimes next week, but definitely not next quarter or next year.

This is essential to create the integrated view of business performance as well as ensuring the impact of tactical decisions are measurable in a timely and accurate manner. In other words, a BI decision support system that is directly coupled to the operational systems.

To give a rapidly growing business its best shot at transitioning from “reporting” into a true BI culture, the BI team needs a degree of independence from mainstream IT activities. It may require its own budget & project planning and should report directly into the CxO and / or board of directors. Its role is to serve the entire business management structure directly involved in critical decision making. It is not just another IT system.

So whilst all the buzz is about big data and the cloud, it’s important to keep our eyes on the main goal. Delivery of timely and accurate data into the warehouse, to support business decision making. The new options can enable benefits not possible a couple of years ago, if they are adopted with a coherent business strategy.

I was fortunate to be indoctrinated into the ‘data centric’ view of IT very early in my career. But it has been a backseat for a long time. More of an ‘engine room’ gig. I think in time, the legacy that the current focus on big data and cloud infrastructure movement will deliver, is the re-focusing of the industry onto that most valuable of business assets. Its ‘data’.




Tags: , , , , , , , , , , , ,

Comments are closed.