Peak 5 Supply Chain
  • Home
  • Services
  • About Us
  • Contact Us
  • Blog

Confessions of a Microsoft Access Junkie

6/28/2014

0 Comments

 
Picture
I don’t really know when I started using Microsoft Access.  It was sometime after IMS, DB2 and Paradox and after I moved from programming roles into modeling roles.  I’ve often said that Access is an excellent tool for the type of work we do as supply chain designers.  It’s easy to use, visual, requires no programming and is easy to obtain.  Regardless of the network design application you use, there’s considerable data manipulation in a network or transportation design study.  In fact, I estimate that on a typical study, perhaps 80% of the time is spent manipulating data outside of the optimization application.  Many commercial-size supply chain design models are small enough to work within Access, so gaining the horsepower of SQL Server is often not necessary.  However, SQL Server can handle larger models and is more robust.

After several tries that led to basic familiarity with SQL Server, I finally bit the bullet and learned it.  Really learned it – so that I can accomplish the same tasks in SQL Server as I would in Access.  I’m talking here about the desktop version of Microsoft SQL Server, also known as SQL Server Express and specifically, the query language T-SQL.  It’s free and can handle 5x the database size that Access can.  This will cover you for most any commercial supply chain model.  One thing that helped my learning a great deal was a book called Microsoft SQL Server 2012 T-SQL Fundamentals by Itzik Ben-Gan.  I highly recommend picking it up if this blog resonates with you.  It’s available on Kindle as well as traditional format.

The disclaimer here is that I have a long programming history, having started my career coding COBOL programs at Exxon before moving into the operations research arena.  Over the years, I’ve continued to write programs in many different languages to help expedite my work or do tasks that queries alone can’t (or require queries that become so convoluted that it’s not worth developing them).  I’ve also seen a good deal of SQL code in passing and understood the concept pretty well going in.  This made it easy to read Ben-Gan’s book and come away ready to use what I learned right away.

Here are some reasons I resisted SQL Server so long:

You just install Microsoft Office and you have Access.  SQL Server is NOT like that.  It’s a number of different applications, server setup, parameter settings, etc.  It is absolutely more like an IT/database application than a desktop tool.

  • Access is visual.  You drag your tables into your queries, drag the fields into columns you want to see, and even drag fields around to create joins.
  • Access queries require no programming.  To use SQL Server, you have to learn to write T-SQL code.  Why this scared me away, I’m not sure.  I guess I just didn’t feel like learning something new when Access had done the job successfully for so many years.
  • Queries can be easily strung together in macros to automate the running of several in sequence.

I finally learned SQL Server because I was embarking on a project that I believed would require a large model and also I began to feel left in the dust by people who did know it.  No 20-something modeler is going to make me look out-of-date!

So I went into learning SQL Server resigned but reluctant.  And I came away… a FAN!  Here are some reasons why:

  • Access queries do one thing at a time.  It’s pretty difficult to see the flow of a set of queries used to populate a model table.  Yes, you can create macros and thus see the flow one query to the next, but read on…
  • SQL Server queries use a file-based approach for invoking queries and other database tasks so you can encapsulate multiple queries that perform a modeling task together, with comments that make the process clear.  A SQL script is simply a text file with comments, commands, etc.
          - “This query creates a list of unique customer IDs”
          - Query listed…
          - “Use the Customer ID list and shipment history to create the Demand table”
          -Query listed…
          -Etc.
  • I believe that once you get used to it, this approach is faster than Access’ window approach
  • Given how you can organize and comment the task flow, I think it’s far easier to come back to the same process a few weeks later, visualize the flow and remember what you did and why.  In Access, this requires opening many windows (queries, macros, field builders, etc.), which is pretty slow if you think about it
  • You can invoke programming/functions for repetitive tasks.  These functions are stored in the database and always ready for use

Still, there are some good reasons to stick with Access for supply chain & transportation network design and they go back to my original reasons for sticking to it so long.  If your model/process will be used or supported by people without programming background and not in a position to take on programming, then stick with Access.

My recommendation is that if you’ve been contemplating but avoiding embracing SQL Server, go ahead and take the plunge.  With just a bit of investment, you’ll be efficient with it in no time.  In fact, I think you’ll be more efficient than with Access.


0 Comments

International Business Travel

11/21/2013

1 Comment

 
I’m just back from a business trip to Norway and have a very non-technical viewpoint to share today.  : )

Over nearly 25 years in business, I’ve had the opportunity to travel to some interesting international destinations.  I have visited all the continents except Antarctica and I don’t plan to go there anytime soon for supply chain design work!

These years overlap with the years I’ve been raising a family, so most international trips were scheduled to minimize time away from home.  Even when I wanted to take a few extra days, it always seemed to be someone’s birthday, there was a playoff hockey game, or something else the prevented an extended stay.  For many years, people would tell me how cool it was that I travelled overseas.  My standard reply was “I’ve been to some very interesting places and saw nothing but the hotel and the inside of an office building, which looks pretty similar worldwide”.

But over the last couple of years, my attitude on international business travel has changed.  While my standard reply has some truth to it, I’ve realized that working & interacting with people from other countries in their normal environment gives you a view on life in those countries that no tourist excursion can provide.  I haven’t seen all of the tourist sites in the places I’ve visited, but I have learned about the country & the people – what they’re like, how they do business, what their families are like, what they think is funny, what they do on weekends, how the politics work, what holidays they celebrate, what they eat (and don’t, and why), how they get to work, how they like their coffee, etc.  I’ve learned about local customs in the workplace, like offering fresh waffles every Friday afternoon – what’s not to like about that?!

With all this in mind, here are a few ideas to pass along…

  • Learn how to say “please” and “thank you” in the local language.  I think people appreciate your attempt.  An interesting phenomenon here is that when I say “takk” or “merci” or the “thank you” equivalent in whatever language, the person typically says “you’re welcome”.  I’ve noticed this in many countries and it amuses me every time.  

  • While you’re there, GET OUT!  Do not eat Marriott room service every night.  Don’t stick with TGI Friday’s (sorry, Jason).  Try the local and/or favorite foods.  If they don’t have an English menu, someone will generally translate for you.  In this case, don’t be picky!  Don’t make them spend 15 minutes trying to explain ingredients in English.  Just pick something that sounds interesting and go with it.  Cold salmon with potato salad.  Sounds great.  Done.  (this is one of my favorites in Sweden)

  • One of my favorite things to do during international travel is “Running as Sightseeing”.   I will take a good look at Google maps (obviously after inquiring about safety and in some countries, acceptability of a woman alone in public), find an interesting park or a street or path and go for a run.  I run with my iPhone, stopping to take pictures whenever I see something interesting.  Sunday morning I had about an hour to spare in Bergen, Norway before my train.  I seriously considered snuggling into my couch and working on my weekly project update.  Instead, I hauled myself up & out and had an awesome sightseeing excursion.  I ran up the winding streets of the high fjord walls and experienced beautiful views over the city, as well as everyday houses that you or I might live in.  I could have glimpsed some of this on a tourist excursion, but even better that I did it on my own two feet.

So next time you travel internationally, get home soon, but while you’re there, do whatever you can to experience life in that country!


(below: snapshots of Bergen, on the west coast of Norway)
Picture
1 Comment

Why Not Use Excel for Supply Chain Network Design?

11/5/2013

0 Comments

 
At the recent CSCMP Global Conference in Denver, the 2nd day ended with a panel on network design.  The 30 or so attendees were approximately split between network design professionals and companies who might need their services. 

Today, I thought I'd elaborate on an interesting topic posed by someone from a company young in network design.  She asked "Why do I need a fancy network design tool?  I have some SQL-savvy guys and they have Excel."  Some network design professionals may scoff at this question, but I think this is a great topic to explore. 

So why should a network design team grow into using a network design tool instead of using Excel spreadsheets?  Well, maybe you shouldn't.  I know network design pros that don't believe in using commercial applications.  They feel that a combination of Excel and some add-on tools for geocoding, distance calculations, maybe a linear programming solver, etc. is more flexible and quicker than a commercial tool.  I wouldn't go this far myself, but there is a time and place for Excel analysis. 

Several years ago, I built an Excel model and associated Mappoint project to determine the optimal location of local delivery centers for a major metro area.  I was able to try various combinations of 1-2 delivery centers using my Excel model to evaluate the trade-offs between delivery cost and real estate cost.  For example, is it better to site the delivery center in a low-rent real estate neighborhood and have more "windshield" time, or in the transportation-optimal, but higher-cost real estate area?  As long as the need was for 1-3 local delivery centers, the Excel model was adequate.

The benefits of Excel modeling are:

  • Anyone who does analysis - financial, supply chain, transportation - knows how to use Excel and is equipped to build a model
  • Excel models are quick to build and unless they use obscure Excel functionality, anyone can understand them
  • Nearly every employee of every significant company has Excel on their desktop.  Commercial network design tools can cost $100k or more to purchase.
  • If you have a small number of on/off decisions (1-2, maximum 3), Excel may be your quickest, most practical choice.

So what's not to like about Excel?  Excel breaks down in some areas:

  • There's an art to building Excel models and there are no standards.  Give 5 people a business model to build and you’re likely to get 5 completely different Excel models
  • Building and evaluating “what-if” scenarios can be challenging because it's difficult to build a scenario framework in Excel.  You typically end up having multiple sheets or workbooks to represent different scenarios.  Consider how hard it will be to maintain 15 nearly (but not exactly!) identical Excel models, one for each “what-if” scenario.
  • But the following are two significant reasons you may want to consider moving to a commercial network design tool at some point.
  • Excel models do not easily allow you to consider facility throughput capacity.  Let's consider that local deliver center example again.  We can pretty easily build an Excel model for a metro area and give the model the ability to choose the closest delivery center for each customer.  Sum up the customer deliveries for each delivery center and you have the projected facility throughput.  Uh oh... what happens if the sum for one of the facilities is greater than what the existing or potential facility can handle?  Play around with this a bit if you don't believe me, but in Excel, it becomes - choose your term - challenging or painful - to try to model capacity on facilities. "For this customer, choose the closest delivery center, except when delivery center is over capacity, in which case consider whether this customer is closer to the delivery center than other customers..." …you get the picture.  You are out of Excel's league for most capacity-related analyses.
  • Excel is DESCRIPTIVE.  Commercial network design tools are PRESCRIPTIVE.  Excel will describe what the output will be given a set of inputs.  Commercial network design tools will take your input costs, network structure and constraints, and recommend the optimal product flow path, facility locations, and other network opportunities. This is where commercial network design applications can really drive value.  Because they're based on a math model of the network, their optimization isn't dependent on people's ideas, which can be limited by the way things have always been done.

Next blog entry, I'll dig into this last point in some more detail...



Also posted as a guest blogger at UltraShipTMS Supply Chain Collaborator blog.

0 Comments

Why Do Transportation Network Design as Part of a TMS Implementation? (written as a guest blogger at UltraShip TMS Supply Chain Collaborator)

10/14/2013

0 Comments

 
Implementing a TMS is great step in improving your transportation performance and has the potential for considerable transportation cost savings.  So why should you consider changes to your network design in addition to implementing a TMS?
Picture
First, let's clarify just what network design is.  Network design optimizes your transportation network to achieve cost savings while maintaining acceptable service levels.  It may involve adding or removing distribution centers, cross-docks or pooling locations.  Network design changes can also involve a shift toward using a different mode, such as rail, or routinely combining LTL shipments into viable truckload shipments.  It might even entail changing production locations for specific product lines -sometimes even to the extent of moving a production line!

Optimizing may be as simple as considering different "what-if" scenarios by building an Excel spreadsheet model.  For example, you may capture inbound and outbound costs and use these along with simple distance calculations to consider an alternative or additional distribution location.  Network design can also involve using a commercial software tool to build a robust model of the network and then exercising this model over a number of different “what-if” scenarios to identify the optimal cost/service trade-off.  Most companies start their network design endeavors with Excel and grow into use of a commercial software tool as their needs and capabilities grow. 

Network design changes can be costly and significant.  Opening a new distribution center or cross-dock is a substantial change not to be taken lightly.  But this type of change can often result in substantial cost savings beyond what a TMS implementation can yield.  All transportation networks change over time - through organic or population growth, acquisitions, new product introductions and other changes.  Oftentimes, operations continue in the same manner, even though the business has changed.  Network design studies give you the opportunity to take a holistic view of the transportation network.  Changes to the design of your network can yield 5-15% savings, depending on the amount of change and the amount of time that's passed since the last network design study.  A TMS system will yield cost savings on a daily basis as planning efficiency improves.  An improved network design has the potential to make a step-wise change in transportation cost that can be realized in addition to daily planning efficiency.

0 Comments

Welcome to the Peak 5 Supply Chain blog!

10/11/2013

0 Comments

 
I hope you will find the topics here timely and interesting.  There will be educational topics, hints & tips, mini-case studies and other supply chain & transportation network design topics that seem interesting.

If you have a specific topic idea, maybe expanding on a blog entry or taking it in a different direction, please let me know and I’ll try to accommodate.
0 Comments

    Author

    Sara Curtis, Principal Consultant at Peak 5 Supply Chain
    learn more...

    Archives

    June 2014
    November 2013
    October 2013

    Categories

    All
    Network Optimization
    Supply Chain
    Supply Chain Design
    Tms
    Transportation

    RSS Feed

Web Hosting by FatCow