Is the juice worth the squeeze?


I have heard this phrase a lot in recent months in a business context.  It’s so visual, I love it! 

It’s not quite enough though.  It’s pretty simple to understand that every project must be able to pay for itself and deliver a return.  Is the juice worth the squeeze?

It’s also true though that no organization has infinite resources of time or money.  If you have 10 projects that you could do but only enough resources to handle 3, you must prioritize those projects that help you meet your objectives (growth, profitability, market share).  What has the most bang for the buck? 

So with these 2 phrases in mind, it should now be easy…right? (Can you hear the sarcasm)?


In the past, I have been guilty of figuring out something can be done and then wanting to rush ahead to do it.  I have been there, I have done it and I have been surprised when a manager along the way was disinterested in my project or actively trying to shut it down. 

My enthusiasm was high, I had the capability to deliver but my ability to see the bigger picture from a business perspective was lacking.  Perhaps it’s a gap in my formal educational process that focused on business math and statistics rather than finance or managerial common sense?  I’m afraid I am not alone however; many analysts and process improvement folks seem to suffer from the same condition.

I think it’s fair to say that my working experience has largely corrected this: a spell in finance working on investment appraisal helped; running a logistics development team that generated more good ideas than we had capacity to work on helped too.  Being responsible for delivering financial results may have been the clincher. 

Analysts can be blinkered, I admit it, but this inability to see the big picture is not restricted to technical folks.  Managers see the complexity of analytics and have their own knee-jerk reactions.  Some perceive high-risk and immediately, perhaps unconsciously, discount the net benefit they are likely to receive.  In other cases I have seen an almost cult-like view that is without justification and disconnected from the results that could have been predicted  (“do this and good stuff will happen”).

In both cases managers are limited by their inability to estimate cost and/or benefit.  This is where a good analyst can really help. 

Note that for prioritizing most projects we do not need extreme precision in cost or benefit.  Really good projects do not have to scrape a return, the poor ones are usually struggling to hit whatever hurdles your finance team has put in place (e.g. NPV, IRR or payback).  What you need is a reasonable estimation backed up by sound analytics and whatever benchmarks you can lay your hands on.  Let’s take a few examples:

Project A:  Implementing an upgrade to the warehouse management system that converts all current paper-based processes to run on the existing computers. 
·         Based off a time-study, this is likely to save 10 minutes per order
·         The network of distribution centers process approximately 3000 orders per day
·         Cost to implement is estimated at 13 weeks of development time.
·         Warehouse labor costs about $25/worked-hour including all benefits
·         Development labor costs around $100/worked-hour including all benefits

Annual Savings:
10 x 3000 x 365 =  10,950,000 minutes/year
= 182,500 hours / year
= $ 4,562,500  / year         
            One-off Costs:
13 x 40 x 100 =  $52,000 / year
            Summary:
Without calculating NPV or IRR or Payback, I think we can clearly see that this would be a very, very good project.  Focus on the savings per person (forgetting that there are a lot of them and you can easily miss finding this opportunity)


Project B:  Report Automation: in our sales office, our analysts currently spend around 10 hours each, every week, preparing standard reports from Point of Sale (POS) data.  Automating these reports would be very popular, removing a tedious, repetitive part of the work.  Automation of each report takes about 1week of developer time. 
·         We have 10 sales analysts producing 40 reports
·         Sales analysts typically cost about $70/worked-hour including all benefits
·         Development labor costs around $100/worked-hour including all benefits
Annual Savings:
10 x 10 x 52 x 70                =  $364,000
One-off Costs:
40 * 40 * 100                       =  $160,000
             Summary:
Our annual savings do outweigh the costs… or do they?  For the savings to be real we have to stop paying for these hours (the equivalent of 2.5 people) or be able to reinvest them into other work that also generates a return.  Will we?  Frankly report automation is more reasonably justified by eradication of error and consistency of output that makes it easier to manage the thing you are reporting on – perhaps $Billions in sales.

Project C:  Use Point of Sale (POS) data to improve the forecast accuracy of the forecast we build for manufacturing planning.
·         Our current Forecast Accuracy is 75% for 1 month out.
·         We believe that incorporating POS data into the forecasting process could improve forecast accuracy.

(This is where your analyst should help, because you really need a lot more information, knowledge of how inventory buffers uncertainty, a decent model, a pilot and good benchmarks to figure out what this is worth)

A small pilot project using POS data and shipment history from the last 3 years to predict sales for last year suggests we could improve forecast accuracy by 3 – 7 percentage points.
Finished goods inventory is what buffers the manufacturing plant from uncertainty in demand.  With a better forecast you need less safety stock (see [How much inventory do you really need ?] for more details and a handy inventory model).  Using the inventory model:
·         The safety stock portion of our overall inventory is currently 1.8 weeks of supply.
·         A 5 percentage point improvement in forecast accuracy (from 75% to 80%) is worth about 0.4 weeks of supply.
·         From Finance we understand that 1 week of supply is worth approx. $12 million at cost.
·         Our weighted average cost of capital is 12% so projected working capital savings are ~ $1.4 million
·         We may be able to save on storage costs too (assuming they are variable not fixed).  Converting inventory in storage pallet positions we estimate saving about 20,000 pallet positions at a current cost of $5 per pallet per month.
(20,000 x 5 x 12)  = $1.2 million
·         Note: there are no ongoing savings to handling costs as we have reduced inventory not throughput (or sales would have dropped too).  A one-off saving in handling while inventory levels fall could be included but would be relatively immaterial.

Our pilot project has also helped us understand exactly how we can enhance the forecasting process with POS data and allowed us to cost the necessary changes to the forecasting system at approx. $1 million in one-off cost.   So we end up as follows:
Annual Savings:
$1.4 million in cost of working capital
$1.2 million in variable storage costs
One-off Costs:
$1 million
Summary:
This juice is (probably) worth the squeeze.  With a payback around 2.5 years it should be on our list of viable candidates.  Remember that the pilot said that accuracy improvement was in a range of 3 to 7 percentage points.  We evaluated the average here. At 3 points the costs will stay the same, the savings would only be 60% (not so good).
Does it give the best bang for the buck?  Well we will have to line it up against all other projects competing for our resources to know that.  My guess… probably not.
By the way, the inventory modeling exercise also said that you have 0.5 weeks of unnecessary inventory in the system.  Perhaps it would be better to start by trying to eradicate that.

The bottom line for you is that you should consider using this sort of Analytical Estimation in deciding which of your projects make the cut.  A good estimate is a lot better than a bad guess.

BTW - from recent experience, I can confirm that beetroot juice is most definitely not worth the squeeze J