Predictive Analytics, Business Intelligence, and Strategy Management

9 12 2009

I was having a discussion with one of my clients this week and I thought he did a nice job summing up Predicative Analytics.

So in the World According to Reed (WOTR) – “queries answer questions, analytics creates questions.” My response was “and Strategy Management helps us to focus on which questions to answer.”

Reed Blalock is exactly right, traditional BI is about answering the questions we know. Analytics is really what we create with data mining – we look for nuances, things that might give us new insight into old problems. We use human intellect to explore and test. And yes, there is a little overlap. But what is really happening is that we have a different level of human interaction with the data.

BI is about history, analytics attempts to get us to think, to change, and idealistically to act.

The danger with both of these is that they can be resource intensive. Neither tool, or mindset should be left to their own devices. What is needed is a filter to identify the priority and purpose. This is where strategy management and scorecarding comes into play. We have built out massive informational assets without understanding where, when, and how to use it. We have pushed out enormous reporting structures and said “it’s all there, you can find anything you need” yet we scratch our heads when we see adoptions levels are low.

What we have typically not done all that well is build out that informational asset by how it helps us be more productive along product lines, divisions, sales region, etc. We have treated all dimensionality the same. Why, because it was easy. The BI tools are tremendous in how quickly you can add any and all dimensions.

“But because you can, doesn’t mean you should”

As we built out these data assets, we did not align them to performance themes.  We have gotten better with some key themes, like supply chain management, and human resource management, but what about customer performance?  We might look at sales performance, but that is a completely different lens than customer performance.

How do we determine which assets to start with…what assets do we need to be successful 3-5 years from now, or what are our biggest gaps to close today.  Think about customer value, or employee satisfaction (and that doesn’t mean more HR assets).  Think about your gaps in Strategy.

How often do we discuss…

  • Are our customers buying more or less frequently?
  • What are our best, and better customers doing?
  • What are the costs associated with serving our least profitable customers?
  • Where are our biggest holes in understanding?
Advertisements




KPI: Overhead per Customer

22 07 2009

If you are trying to measure management improvement, how about looking at Overhead per Customer (or per transaction).  This should be a decent indicator in terms of management and overhead scalability.  If we are doing a better job of managing the business we should see some increased returns in the management function.

  • If the trend is increasing, we should be discussing the scalability of the organization.
  • If the trend is decline, is it for the right reasons?

While you are at it, you might also include cost of sales per transaction.  This one is perhaps a little more debateable in that we don’t want to artificially manage this number.  Reducing the number of sales reps, may drive down the number.  Reducing compensation plans may chase away our better sales reps.





Interesting Financial KPIs (REL Consulting)

28 05 2009

I found a few KPIs on the web that were worth discussion.  These are from REL Consulting (under solutions > key metrics).  These are both great KPIs for the CFO as well as for the organization to better understand.  

Here are a couple of initial thoughts:

  • As these are really Financial KPIs, which while clearly quite important, don’t always make good operational indicators.  Financial KPIs only make up part of the overall health of the organization.  
  • These are also lagging indicators, as are most financial metrics.  By the time these have changed, chances are we have already been through some business cycle and have potentially lost value we could have identified earlier.
  • The trend here is clearly important, as a change in any of these is clearly an alarm for their business.  These might be more monitoring type of metrics (sorry Jonathan, had to throw it in here), if the trend were to change it might indicate further analysis to define specific actions.  
  • We need to make sure we are not managing to these artificially.  Take for example, Days Sales Outstanding (DSO), this is often a proxy guage for customer satisfaction.  If marketing/sales changes payment terms just to impact this number we are doing the wrong thing (potentially).
  • Perhaps the greatest strength of these KPIs are as teaching tools.  If we are saying this is what is important as an end result, then we need to be able to communicate what they mean and how each of the individual processes, programs, and initiatives impact these metrics.  Most people in the organization don’t understand the financial outcomes with exception to revenue and perhaps margin.  HR and Finance should make this part of their training regimens to improve overall financial knowledge.
  • We also need to understand how these impact strategy.  If we are a high tech company with a priority on market share, chances are a lot of these numbers are going to slip.  Perhaps the converse is more interesting, if the market wants us to focus on market share growth, we need to be careful that these are not sacrificed to an inappropriate level.

Again, I am not saying these are anything other than great measures.  Nor am I picking upon REL as I have heard very good things about them.  But as you are designing KPIs, they need to be easy to understand, linked to strategy, balanced across functionality, and more often than not leading indicators.

Best Possible Days Sales Outstanding (BPDSO)

BPDSO is the value that achieved if all customers paid exactly to the agreed upon payment terms. Typically a business will offer more than one payment term to its customers and therefore the BPDSO takes the different payment terms offered into consideration by using a weighted average based on value of sales/revenue by payment term. This measure is often called the theoretical days sales outstanding (DSO) because in reality it is almost impossible to actually achieve as there will always be customers who pay late and other external factors hindering receipt of payments to term (e.g. banking delays, post service delays, etc.).

Cash Conversion Efficiency (CCE)

CCE looks at how efficient companies are at generating free cash flow from operations, or operating cash flow from sales revenues – how much free cash flow makes the journey through the operating cost structure of a company. While CCE is a simple metric to derive using it can provide powerful insights into the overall health of an organisation’s cash-generation capabilities.

Days Inventory Outstanding (DIO)

DIO is financial and operational measure, which expresses the value of inventory in days of cost of goods sold. It represents how much inventory an organisation has tied up across its supply chain or more simply – how long it takes to convert inventory into sales. This measure can be aggregated for all inventories or broken down into days of raw material, work in progress and finished goods. This measure is normally produced monthly.

Days Payables Outstanding (DPO)

DPO is a relative measure of a business’ outstanding payment liability. DPO measures the level of outstanding payments at the end of a month expressed in terms of the number of days payments represented by the creditor balance, i.e. the number of day’s worth of payments still outstanding. The metric is useful as it gives an indicator over time of what payment terms are being accepted and complied with within a company.

Days Sales Outstanding (DSO)

DSO is a relative measure of a business’ debtor exposure. It measures the level of outstanding sales/revenue at the end of a month expressed in terms of the number of days sales/revenue represented by the balance of the accounts receivables (i.e., the number of days worth of sales/revenue still outstanding). This measure is typically represented as a monthly trend and is important as the increase in the gap between DSO and BPDSO can be an early sign of deficiencies in the credit and collections process. When determining if the DSO of a company represents good performance, it should be compared to the company’s BPDSO. BPDSO is important as a reference point against which to compare a company’s DSO performance. A DSO of 92 may initially appear to be very high, but if the company’s BPDSO is 88, then a DSO of 92 represents a good performance.

Days Working Capital (DWC)

DWC is a measure of the cash conversion cycle that gives insight about the underlying health of a business. It is a key metric because it measures the average number of days of tied up working capital in the operating cycle. If DWC is trending upwards over time then it will have a negative financial impact on overall company profit.

Forecast Accuracy (FA)

Compares the ratio of forecast error to actual sales and is expressed in percentage terms. It shows the accuracy of the sales forecast compared to actual sales within a period of time, normally a month. Forecast accuracy typically shows better results when we are predicting demand for the next weeks as opposed to the next months, or where we aggregate the measure for a group of items. The more accurate the forecast, the easier it is to manage inventory levels across the supply chain.

Return on Capital Employed (ROCE)

ROCE is a ratio that indicates the efficiency and profitability of a company’s capital investments. The measure is important as ROCE ratio should always be higher than the rate at which the company borrows, otherwise any increase in borrowing will reduce shareholders’ earnings.

Shareholder Value Add (SVA)

SVA is a value-based performance measure of a company’s worth to shareholders. The basic calculation is net operating profit after tax (NOPAT) minus the cost of capital from the issuance of debt and equity, based on the company’s weighted average cost of capital. All working capital improvements help improve SVA.





Manage vs. Monitor

21 05 2009

It has always struck me as a little odd that a great deal of marketing literature in the Business Intelligence and Performance Management space talks about “monitoring” performance.  Isn’t the entire goal of this space to help companies actively “manage” their business.  My concern is that “monitoring” assumes that all is well unless some alarm is triggered.  

While it is fine for the thermastat to monitor temperature, perhaps business is a tad more complex.  Instead of waiting for things to get to a threshhold, we need to understand a number of things that all work more or less together to explain a more complex concept.  

Instead of just showing up for a meeting, what if we focus on creating a culture of being prepared for a meeting.  We can then use Business Intelligence as a organized and focused set of tools to help with our prep work.





Relevance and Context

20 05 2009

With times a little tight these days, BI projects need to be more focused.  While there are a number of ways to do this, starting with very specific briefing books targeted at a management process or a departmental data mart, think about using relevance and context.  Don’t just recreate the hundreds and thousands of reports that have been created before, use new money to rethink old ways.

For example, compare a Google search versus the WolframAlpha search engine.  Reporting environments can often look like a Google search list – while there is some relationship, there may not be much relevance.   I may have to search around quite a bit to find what I need.  Enter the WolframAlpha search, it requires the user to provide appropriate context.  With the right context, Walfram works wonderfully, without it, not so much.  It only does what it was designed to do.

Like BI tools, each search engine is designed to do different things.   By training the users to use the right tool with the right context you have a greater chance to provide people with the information they need to make decisions.  You would not ask the same business questions to cubes, reports, dashboards, scorecards, etc.

 

And yes, perhaps I did force the argument so I could say “WolframAlpha.”  And with that said I should probably give a shout out to the folks at Cuil.  





Scorecarding – Getting Started

18 05 2009

Scorecarding, or Strategy Management, is a journey.  It is more important we get started and learn and adapt as we go.  One reason why scorecard projects stall is that organizations expect immediate maturity.  It takes time to understand the different stages, and the different stages are important and valuable points of learning.  

  • Start small and focused with a team that has a well defined management process.
  • Don’t make changes every month, give the concept a quarter to learn.  Then meet to make the changes.
  • Use the concept to facilitate conversations about what creates value.




The Value of Scorecarding

21 04 2009

One of my first Scorecard exercises is one of my favorites.  It taught me a great deal about the power of scorecarding.  I did what I suspect most people do.  I interviewed all the VPs and developed a long list of KPIs.  I then used an excel spreadsheet to organize the KPIs.  I put the KPIs down the rows, and theVPs across the columns.  Then to help visualize the data, I placed “red” cells where VPs were directly impacted by the KPIs and “yellow” cells where the VPs were indirectly related.  I did not intend the colors for anything other to call out attention for each of the VPs.

By choosing the “red” and “yellow” I had each of the VPs concerned that they were under performing in each of those areas.  I had to explain a number of times, the reason for the colors.  
  • The first lesson was that by associating colors with performance, I clearly had the attention and focus of the executives of this team.  It sparked a number of very strong conversations about performance.
  • The second lesson is that communication is just as important.  By doing a less than stellar job of communicating (at least from a visual sense) the information, I wasted a tremendous amount of time that should have been used for more strategic discussion.  

Scorecarding can be a very powerful tool, but it needs to be used appropriately.