From Data to Action
I (Ines Cibrian, Sweetspot Intelligence) recently had the opportunity to participate in the 3rd edition of MeasureCamp, where among other interesting questions, someone came up with this question: “What do you do to SHARE the information in your DASHBOARDS?” I thought this could be answered quite comprehensively with a little time, and today I have that time. So here we go.
It is no doubt important to count on a regularly updated dashboard that remains tied to your business objectives. But sharing the insights that stem from exploring those results is often more important: Collective action plays a key role in the improvement of outcomes over time.
Now, how are we ever to achieve this considering that:
- Each stakeholder is better served by a highly-customised dashboard that speaks her own language (what better way to meet our own specific objectives?), meaning, no two individuals end up receiving the same information.
- The nature of the insights required by each level in the organisation will also vary: Whereas operational layers expect tactical proposals (such as campaign micro-management tips), strategical ones long for data revelations that will help them steer pricing policies or customer care investments in the right direction.
And the answer to this lies in the cascading structure of optimal decision-making environments. In other words, we should seek a relationship between metric groups that imitate the relationships existing between the actual departments involved. This will ensure that every micro-action taken at a lower level in the organisation ends up having an impact on the KPIs consumed at a higher level.
Of course, once this hurdle is out of the way we are faced with the one issue that I have come here to discuss: How do we now share insights (valuable, actionable information), when all we have is a reporting delivery ecosystem (through interconnected, personalised dashboards)?
The answer to this lies in the process that has led to the acquisition of such insights. Such process would normally follow one of these two patterns:
- The analyst, having a good sense for the primary goals of her primary stakeholders, will proactively come up with insights and recommendations within a few days after regular outcomes become available (say, on the 5th of every month).
- Stakeholders themselves will request an explanation for the outcomes received and it is this request that will trigger the analysts’ quest for an answer and/or potential solution.
Scenario 1 can be easily solved through the drill-down features of some dashboards: A KPI or table detail would be accompanied but such explanations or recommendations.
Here you can see an example of scenario 1.
Scenario 2 is slightly more complex, as it already requires the setup of a basic workflow. Such workflow would work best when fully aligned with the dashboard/stakeholders relationships discussed above.
A simple way to articulate such workflow (and communication process) could be structured as follows:
- Business users access their regular outcomes (via their personalised dashboard). They may notice that one of their KPIs is doing badly or just not meeting their goal. Such KPI is highlighted and a comment is attached to it.
- Analysts receive an alert (via email or through the dashboard’s own alerts), go back to their analysis tools, slice and dice their data and come back with a “Why”.
- Business users may not agree with such explanation and request a more convincing one, or they may like it and ask the analyst to “go for it” in the form of a “How to” proposal.
- Analysts build a hypothesis (the “How”) based on the data at hand, possibly building a prediction for future potential outcomes.
These action-driven reporting scenarios are already well covered by the Digital Insight Management discipline, and you may read more about it (and access a couple of real case-studies) from Digital Analytics old timer Eric Peterson, who last year published a white paper about it available at http://www.sweetspotintelligence.com.
INÉS CIBRIÁN, Marketing Executive, Sweetspot Intelligence