I was recently sent a copy of a Gleanster report. Gleanster for those who have not heard of it is a new company (founded by an old colleague of mine, Jeff Zabin, and others) that delivers free research. It is not clear what its business model, exactly, but for readers of the blog it represents a great source of free research.
I was recently sent a copy of a Gleanster report. Gleanster for those who have not heard of it is a new company (founded by an old colleague of mine, Jeff Zabin, and others) that delivers free research. It is not clear what its business model, exactly, but for readers of the blog it represents a great source of free research.
The report was on Business Intelligence (you can get it here http://bit.ly/f9I9Vw) and while it covers mostly what I would consider “traditional” BI rather than the kind of operational analytics on which I focus, there were some good nuggets and interesting observations. In fact it began well, by pointing out that data mining and analytic modeling should be part of a business intelligence strategy even if they are not part of your Business Intelligence software stack. However the report’s focus remained on decision-makers and delivering information to them (decision support) with little or no discussion of the role of analytics in delivering insight to systems (decision automation) – see this blog post on the differences between decision support and decision management.
The report divides up into a number of sections, which look to be standard sections in Gleansights as they are called. My comments on each section are embedded:
- Topic Overview
I would like to have seen some discussion of the role of analytics in information systems (decision automation) as well as in decision support. A reasonable summary. - Reasons to implement
I liked this section as it took compelling reasons to implement given by those surveyed and prioritized it based on responses from companies that are high performers on some standard business benchmarks. And top performers clearly see the value of “smarter, more timely business decisions” and focus on upside (“identify new revenue/growth opportunities”) before cost control. I was a little surprised that “increased customer profitability” did not make the top 3 but perhaps this is because it takes data mining and analytics, not just reporting and dashboards. While the clear focus was on insight for decision makers there was a nice comment:
While revolutionary breakthroughs may have to await a flash of inspiration, often the potential for new business is written in the company’s numbers for anyone who takes the trouble to look”
(see my post on the myth of the aha moment for a similar take). - Value Drivers
Again, the approach of focusing on the drivers noted by top performers was nice and it was interesting to note that “customized interfaces” and “foster a culture of data-driven decision making” both scored highly. For most companies I would argue that a focus on operational decisions and on data-driven decision making by front-line staff is also required. Of course for these folks a customized interface is one that simply consumes analytic insight not one that presents it e.g. a call center application where the cross sell displayed is selected analytically. While this section did talk about data mining and predictive analytics, its focus was on the power of embedding these analytic approaches in reports and dashboards and this is not what I see as their primary value. I was surprised and disappointed that continuous improvement scored so poorly. My experience is that no decision is good for ever and so constant Decision Analysis and improvement is a must and should be built in from the beginning of any analytic project. - Challenges
I liked this section. I would encourage those struggling with breaking down silos to begin with the decision in mind and to remember that integration of analytics with CRM/ERP systems is not just about sucking data out – it should also involve pushing analytic insight back into those systems to close the loop. I did like the data quality challenge though – “achieving an acceptable data quality”. Data quality is not an absolute measure – you should aim for data quality good enough to support better decision making. The issue of getting managers to stop making “gut” decisions is highly relevant (and something I have blogged about) but even those managers who like making gut decisions should be worried that their call center staff are doing likewise if analytics are not being embedded in your operational systems. - Performance Metrics
The key metric of time to better decision came through loud and clear but most of the rest were focused on BI implementation success rather than better decision making, which is perhaps only to be expected. - Success Story
A nice, if somewhat vague, case study. - Vendor Landscape
VERY focused on the BI tools space, I liked the pattern of letting the vendors describe themselves and then providing a Gleanster comment on each. Would like to have seen the data mining and predictive analytic vendors getting a look in
Overall I thought the report was well done and, if you were looking for some good tips and ideas for BI implementations, this would be a good report to download and use. Hopefully the folks at Gleanster will follow-up with something on data mining and predictive analytics before too long.