A Data Science Central Community
Taming the analytics hydra with an Analytics Center of Excellence (ACE)
by Scott Mongeau
Managing analytics, organizationally, has the aspect of a hydra, the ancient Greek mythological dragon which Hercules slew. For each head Hercules dispatched, two more burst forth; only by cauterizing the severed heads and focusing on the body, was Hercules able to best the beast. Whether or not the ancient Greeks were dispensing on the problem of project politics, with Hercules as a harried manager, they leave us with a gravid metaphor to apply to organizational initiatives that swiftly grow out of control.
Enter the fevered hype surrounding business analytics: we see that analytics projects are popping-up all over the typical organizational map. When management attempts to marshal and frame the initiatives (i.e. to prevent scope creep, to translate into wider organizational objectives, to implement best practices), two more rogue analytics initiatives pop up elsewhere. In the hydra metaphor, the heads represent focused analytics implementations arising from particular business silos (the necks), whereas the body represents the centrally important aspect of organizational context.
The danger to organizations that allow unrestrained ‘heads’ (grassroots analytics projects) to grow un-pruned is that a confusing, unwieldy set of conflicting decision-making initiates emerges which drag down and encumber the body. When we consider that the heads, as analytics initiatives, are indeed attempts to assert decision-making authority, we see that the more heads that emerge, the more opportunities that arise for conflict, paradox, and conundrum: pandemonium!
Recent research has proposed that there are two central paths to adopting organizational analytics: top down and bottom-up. In particular, the recent MIT Sloan Management Review article ‘Analytics: The Widening Divide’ (Kiron et al, 2011), proposes the two paths as being, respectively: Collaborative, a centralized corporate competency; and Specialized, bottom-up based on a particular business focus such as finance, market, or operational analytics.
The bottom-up or Specialized approach involves growing analytics competency based on a specialized function. In this frame, analytics emerges from a focused business area, typically finance, sales / marketing, or operations / manufacturing (Kiron et al, 2011). The benefit here is of focused, practical expertise gained through meeting hands-on tactical challenges. Typically there is a strong success story in the particular business area which propels the organizational visibility of analytics to the executive level, giving momentum to broader adoption. The danger is that this approach forces ‘one-size-fits-all’ implementations: what works in financial analytics will not fit one-to-one in a sales and marketing environment without translating to different cultures, drivers, and demands. In contradistinction, the danger of the Specialized path is that the momentum is stillborn and loses steam once it is realized that the ‘success story’ is captive to a particular silo and does not translate without a broader, transformational, pan-organizational perspective.
The top-down, or Collaborative, approach is an executive-level sponsored centralized initiative. Although quite high-level, the book ‘Business Intelligence Competency Centers’ (Miller et al, 2006) gives an overview of this centralized approach. In this frame, a general facility of expertise or ‘center of best practices’ is made available and customized to meet specific challenges across the enterprise. Here we propose the term Analytics Center of Excellence (ACE). In this path, IT often takes a centralized ‘stewardship’ role, mentoring particular business silos towards the methods and processes associated with high-quality, analytics-driven decision best practices (Kiron et al, 2011). Centralization allows for the efficiency of ‘getting everyone on the same page’. The challenge is that the central ‘steward’, be that IT or an independent facility, must demonstrate a rare combination of both technical expertise (tool-based, data management, implementation, and methodological fluency) combined with organizational savvy (sensitivity to the often shifting and conflicting demands of diverse expert user needs in multi-stakeholder, matrixed environments – in short, organizational and stakeholder insight).
Both top-down and bottom-up approaches have strengths and weaknesses, but in terms of synergies and efficiency, there is a strong case to be made for establishing an Analytics Center of Excellence (Collaborative approach). In particular, this allows best decision practices to be socialized and standardized, technology to be shared and centrally administrated, and decision challenges to be compared like-to-like in a portfolio environment (much like project portfolio management or centralized IT project management).
This last item, forming a center of decision ‘project oversight’, is subtle, but potentially ‘priceless’ in terms of driving organizational analytics value. Indeed, having a central office for ‘analytics decision competency’ is part-and-parcel of high-risk-high-reward industries already: petroleum majors, biotech firms, and high-tech companies, particularly semi-conductors. In this respect, a centralized office orchestrates decisions, often segmented into projects (indeed such offices are often framed as project portfolio management centers), so that like-to-like value analysis can be implemented.
This often takes the form of translating decisions into cross-comparable risk-reward ratios, otherwise known as Sharpe Ratios. The simple, central argument of the Sharpe Ratio is that in all cases firms should defer to those projects which offer the highest aggregate potential for return at the lowest aggregate proportional risk. In terms of oil exploration or the biotech realm (the biggest users of this methodology), this would mean that those projects that offer a higher proportional potential return in terms of the risk (i.e. volatility of potential NPV outcome upwards and downwards) would rise to the forefront in terms of desirable initiatives. While not all decisions are easily framed as ‘projects’ per se, all business decisions, by nature, should ideally be reducible to a Sharpe Ratio, or, in other words, a measure of aggregate risk cross-referenced with aggregate potential reward.
To the degree that analytics is, these days, omnipresent and of central interest to firms struggling under the weight of complexity, it risks becoming an unmanageable hydra, the multiplicitous heads squabbling and gnawing to have their particular decision insight thrust to the executive fore. The key to taming the hydra lies in adopting a corporate Analytics Center of Excellence (ACE) to mediate between these complementary, but often clashing forces. The challenge is to ensure the ACE is not technically focused, but is framed as a organizational decision steward-cum-orchestrator. This can only be achieved by mentoring an effective mix of organizationally-savvy analytics experts who put business context in the forefront when framing decisions.
Kiron, D., Shockley, R., Kruschwitz, N., Finch, G., & Haydock, M. (2011). Analytics: The Widening Divide. MIT Sloan Management Review (Special Report).
Miller, G. J., Gerlach, S. V., Brautigam, D. (2006). Business Intelligence Competency Centers: A Team Approach to Maximizing Competitive Advantage. New Jersey: John Wiley & Sons, Inc., SAS Business Series.