How do you know your analytics is failing?
How do you know if your company’s analytics is failing? And if you find out, how do you fix the problem? Dysfunctional analytics can either be an obvious or latent condition depending on the situation, but executives should be aware when symptoms of dysfunction emerge and move quickly to deal with it. The cost of mis-informed decisions can be greater than the opportunity loss of having no information to begin with.
One management book I love re-reading is Patrick Lencioni’s Five Dysfunctions Of A Team. The book details five common causes of team dysfunction, the likely behavioral root causes, and ways to deal with them. I’ve personally used the framework in effectively managing small and large teams, and I also find them useful taken in the context of analytics.
Symptoms of dysfunction
- Inattention to Results – Company reports, data and figures are prone to errors, missed deadlines, and failure to tie up with past numbers.
- Avoidance of Accountability – no one is taking responsibility for reporting errors and delays, teams point fingers at each other, lengthy email trails – mostly about assigning blame rather than resolving issues.
- Lack of Commitment – memos and directives on report errors and problems are filed and go unheeded, diagnoses and solutions to the reporting problems are superficial, status quo prevails. Critical issues make a big splash one moment and then die down for a few periods only to get resurrected under a different name.
- Fear of Conflict – despite self-evident flaws in reporting and analytics, everyone acts as if everything is moving along, business-as-usual. Inquiries are often answered by half-declarations veiled in ambiguity (“they’d rather not say”). No root causes to the problem are uncovered. There is an artificial harmony prevalent among the analytics teams. Analytical investigation and data discovery is a minimal to nil practice.
- Absence of Trust – no authority on analytics is mutually recognized by all teams and team members – often defensive of each other’s outputs. Multiple versions of the truth exist and there is either general disagreement about analytics strategy or no strategy at all. Senior management refuses to rely on any analytics in decision making.
- Status and ego – teams or team members do not collaborate with each other in generating analytics and report figures only minding their own patch – leading to inconsistencies in data handling, transformation, preparation which in turn lead to rework and delays. Teams seek to discredit each other’s numbers rather than working together to align them.
- Low standards – responsibility is easily deflected due to low performance standards set by management (if at all). No one feels threatened by bad performance on their part and conversely no one aspires to perform better.
- Ambiguity – with no clear direction, reporting teams are rabidly territorial, there is no buy-in on the ground for decisions made on top and the issues are often considered too complex for any single party to correct. Any queries are bogged down by over-analysis without clear outcomes.
- Artificial harmony – teams prefer the security of harmony over any form of conflict and bona-fide resolution of a core issue. Any remotely incisive query on reporting and analytics output is seen as destructive and teams would rather accept the status quo, no matter how jaded and imperfect, as the way things should always be.
- Invulnerability – teams and managers are more bent on protecting their reputation rather than admit any weakness. Gut-decisions persist even in the face of contradicting data and evidence.
Any management framework like Patrick’s 5 dysfunctions is only as good as the users of the framework. The key is not just proactively spotting the symptoms early enough, but also moving quickly and decisively to address them. Of the symptoms, reporting errors and delays are the first things to spot, but they don’t necessarily denote any real dysfunction since these things will happen over the course of time. What should cause concern is if the problems in the way a company generates information and insight keep coming back or no clear course of action to resolve them seems to be evident. Worst of all: if no one appears to care about it.
Team dysfunctions hinder analytical competitiveness
Continued prevalence of dysfunction may also indicate a problem in how analytics is viewed by the company. In Competing On Analytics, Tom Davenport and Jeanne Harris shared four (4) clear pillars common among analytically competitive companies. If and when symptoms of dysfunctional analytics emerge, I find it useful to start asking questions around these pillars to help spot the underlying root-causes of the problem:
- Senior Management Commitment – is analytics supported by the prevailing culture of senior management? Is fact-based decision-making a priority? Is data gathering and accuracy a key priority ahead of decisions?
- Analytics as a distinctive capability – does the company view its analytics as a key differentiator in its market or against peers or just another management toy to play with? Is analytics aligned to the company’s vision, mission, and objectives? Is analytics viewed as a competitive advantage or a new chore dictated by the management?
- Large-scale ambitions – does the company aspire to dominate its industry through analytics? Does analytics support clear measures which define success for the company?
- Enterprise-wide analytics – is responsibility for analytical output entrusted to all business units, or just the role of one team? Is data and insight intended to be siloed in the hands of few players or made available to all levels of the organization?
Although based on Harris and Davenport’s research, the presence of all four pillars above indicate a company that is analytically competitive – I also believe when one or more of the pillars above do not yet exist in a company – that absence can be the breeding ground for the dysfunctions to occur.
Going back to Lencioni’s framework, there are also five paths to resolving dysfunction. They appear simple, but also taken against the four pillars above, they are very straightforward ways to achieve competitive advantage through analytics:
- Collective Outcomes – teams should be encouraged to share in collective goals to avoid unnecessary siloes and territories. Make the end output of any report the joint responsibility of those extracting, transforming, and preparing the data. Encourage integration and reuse of data whenever possible – rather than teams keeping multiple copies of the same information independently.
- Call out Accountability – apart from the numbers generated by analytics and reports, measures and processes should also be set in place to monitor the analytical outputs themselves like reporting service levels (SLAs) and some basic key performance indicators (KPIs) such as error rates – and teams should be both rewarded for achieving them as well as penalized for repeat failures.
- Clarity of Objectives – establish clear strategies and outcomes not just for the production of analytics but also the rationale for it to exist (i.e. the company’s competitive advantage). Ensure that all teams are aware of this to enforce adoption.
- Constructive Debate – create adequate and ideological tension between teams to keep them on their toes. Encourage members and teams to challenge each other’s reports and data with the aim of improving quality of analytics output, performance, and drive better insights. Data discovery and investigation is encouraged to unearth insights outside of the usual coverage of standard reports.
- Leaders Take Initiative – leadership should take the first steps to expose themselves to criticism, resolve issues, and expose weaknesses in order to generate trust and positive buy-in from all team members and stakeholders. Even low-ranking members are encouraged to speak up to voice out insights, new ways of doing things, and calling out problems as they see it. Leaders also encourage accurate and timely analytics by being the first to drive fact-based decision making not just about company-level decisions but also the way analytics is utilized and delivered.
At the core of it, the symptoms, root causes, and solutions to dysfunctional analytics is the same as that of dysfunctional teams. On the other hand, a healthy team culture also reinforces a healthy analytical culture and vice-versa. As companies move forward in this era of big data they will have more to gain from driving better analytics, and also a lot to lose by not encouraging it.