The Commercial Environment: Pt 2 - The Subtle Deviation
Some things you hear once and they stay with you forever…
"This is the best strategy document I have ever read."
Some things you wish you could remember…
"Honestly, did I leave the front door unlocked all night again? Ah well, it's a safe neighbourhood."
Some things are statistical definitions…
Sample Size (n) - The number of data points you have collected
Expected Value (μ) - The most likely result
Standard Deviation (σ) - How spread out your results are around the expected value.
Standard Error (SE) - how much your sample mean is likely to vary from the expected value, given your sample size.
Sample Mean (‘x hat’) - The average of your observed results over a given period
Confidence Interval (CI) - A range within which the true expected value is likely to fall, given your sample size and standard deviation.
Critical Value (Z*)- The number of standard deviations from the expected value that defines the boundary of your confidence interval.
…to name just a few.
Hold onto them until the end of the post.
* * *
THE COMPARATOR LOOP
In PT1 - An Uncontrolled System, the comparator goes under-examined. For comic effect, I favoured a snide jab at my fellow salespersons over a proper functional description.
I should correct this…
In the minimalist model, the comparator functions qualitatively, taking in the feelings of the salesperson on the basis of the feedback of the outcome. When the outcome is good, the salesperson takes all of the credit, the glory and the bonus.
The opportunity for deeper understanding of the commercial environment is missed.
When outcome is bad, the feedback is refracted through the emotional prism of a salesperson’s disappointment (the most righteous disappointment there is) and possible feelings of threat. The feedback is liable to be lacking objectivity, to be seen as lacking objectivity, or both.
The opportunity for deeper understanding of the commercial environment is obscured.
In the maximalist and hybrid systems the comparator is doing a quantitative assessment, where it measures actual data from outputs, compares it with known data from previous outputs and then concludes on whether the outputs are high, expected or low.
The opportunity for deeper understanding of the commercial environment is omnipresent.
* * *
Some things you discover…
"When I saw the size of the bonus I nearly spat my ginseng all over the place!"
Some things discover you…
*** Ne’er-do-well tries door***
Door: ‘Cccrrreeeeaaaakkk’
. . .
Some things are mathematics…
* * *
“But, what does the
comparator actually do?”
THE MATHSY BIT
Well, firstly, there is a distinction to be drawn in the proposed measurements data.
Some data over any given period can be expected to form a normal distribution and will have an average value, a standard deviation and a clearly defined sample size.
Other data, however, are themselves an individual datum. To generate a normal distribution from these, you would have to measure the same data simultaneously across multiple timelines.
Given that logistical challenge, we will use two separate statistical models instead.
One Sample Z-Test
Above, we have an inequality.
This inequality compares the measured mean to the expected mean, and by selecting a critical value from the table in Fig 7, we can say with a defined degree of confidence after n samples whether the difference between the measured and expected means are statistically significant signal, or background noise, where for larger confidence intervals we need a larger sample size.
Fig 9, below, gives us a more intuitive understanding of the maths,
If the sample mean is as shown in the graph, then we can say with more than 50% confidence that this is a significant change in the expected conditions.
One Sample Poisson Test
Above, we have an inequality.
This inequality compares the measured number of events to the expected number of events and by selecting a critical value from the table in Fig 7, we can say with a defined degree of confidence after n events whether the difference between the measured and expected means are statistically significant signal, or background noise, where for larger confidence intervals we need a larger sample size.
Fig 10, below, gives us a more intuitive understanding of the maths,
If the sample mean is as shown in the graph, then we can say with more than 50% confidence that this is a significant change in the expected conditions.
Comments
The Z-test and P-test sections may look pretty similar (some would say, almost indistinguishable).
There is good reason for this. We are assuming a normal approximation for the P-Test and we are assuming that the event counts from the last comparable fiscal period are the mean.
“That’s a big assumption, Ian, what if last year was an anomaly?”
Good question - remember, we are creating a trigger to further investigation. If there is a significant shift in event counts between expected and measured, then this should be investigated, and only through the investigation would we find if last year was anomalous.
“50% is a really low confidence interval”
Again, good point - but remember what we are doing. Anything above 50% means that it is more likely than not that the difference in mean is due to something structural. This is the earliest possible early warning signal - you exist in a competitive environment, the earlier you understand the commercial environment, the better you can adapt, the stronger your competitive position.
* * *
Some things are used to measure you.
***Swag bag, full of awards, clinks into the night time***
Some things help you to see the truth
***Swag bag, full of plunder, clinks into the night time***
Some things are comparison parameters…
* * *
“Great. But, what does the
comparator actually do?”
The Metrics
With these statistical models, the comparator can assess real time order data from different parameters against the expected values from the last comparable fiscal period with the assumed same commercial environment, which could be last year, the same quarter last year, last quarter or others, depending on your business.
Name: Order Value
Definition: The booked value of each order won
Test: Z-Test
Example Signals: Market is improvement, increasing scarcity of competitive supply (high) decreasing availability of budgets, new entrants in customers market (low)
Name: Order Volume
Definition: The number of orders won
Test: P-Test
Example Signals: More decisions in market, improved coverage, competitor has weakened (high) Market contracting, coverage reducing, or win ratio falling (low).
Name: Sales Cycle Time
Definition: The time elapsed between first meaningful commercial engagement and a buying decision
Test: Z-Test
Example Signals: Market caution, more stakeholders, regulatory uncertainty, or sales engaging earlier (high) More efficient process, customers more decisive, or sales being brought in later with less specification influence (low)
Name: Total Buying Decisions
Definition: The total number of opportunities you were involved in that reached a buying decision
Test: P-Test
Example Signals: Market generating more activity, or your coverage has improved. (high) Change in buying behaviour, budgets squeezed, customers deferring, or sales not getting involved in enough winnable opportunities. (low).
* * *
Some things are falling into place.
Some things are still a little out of sight
Some things are better left for next time…
* * *
In the next post, I will give you a free comparator and show you how to use it.
The opportunity for a deeper understanding of the commercial environment is yours.