Category Archives: Market Intelligence

Outsourcing? Two Years to Turnaround!

While short and sweet and filled with a number of (now) obvious pieces of advice, a recent article over on CIO Insight on “Nine Things No One Ever Told You” [about offshoring] made two great points that many articles miss:

  1. Process Matters and
  2. It Takes Time to Get a Stable, Productive Processoften 18 to 24 months if you and your partner are new to outsourcing!

In other words, if this is your first time, expect that it will take two years to get the full extent of the payback which might not be as much as you expect after infrastructure investments, change management costs, travel, and rework are factored in (which will often be [much] more than you expect).

And process matters. Without sound processes and standards in place to keep the business running smoothly, it is more work to manage multiple relationships, each of which will function poorly without a strong foundation. So, if you still haven’t realized that the outsourcing craze has finally stopped and decide you want to chase after the runaway train long after it has left the station, remember this: Two years to turnaround. Ouch!

Share This on Linked In

There’s No Intelligence Behind a Spreadsheet

Not even of the artificial kind. So please, kill those electronic cockroaches now before the infestation becomes so big that the only way to remove it is with enough C-4 to totally obliterate the entire office building.

So what set off this latest rant? Rich Wilson’s comment at the CPO Agenda roundtable in London this May on “budgeting for a wider influence” where he said:

We developed these powerful analytic capabilities that we have applied to high spend categories, but people weren’t using it. So what we have today is a device called RFxpress, for taking an ordinary Excel spreadsheet that is fed into the front end of our application and configuring it. In essence, it enables the user to have the full power of our analytics to invite suppliers at the push of a button and conduct bids over the internet.

BULLCRAP!

Spreadsheets are NOT an analytics tool. They’re a ledger, which can be used as a poor man’s tool for data capture, but not for manipulation, sharing, or collaboration. Why? For starters:

User Entry Can’t Be Controlled

Sure Excel allows a user to define the type of a field and even lets a user define a few macros to check and format data, but considering that another user’s environment might have macros disabled (and, like Office 2008 on Mac, might not even support VB macros) and that any user can override cell types, a user can literally enter anything they want if they have even one iota of technical proficiency.

Cells and Computations Can’t Truly Be Hidden

A…C…K? Better unhide those columns in case they are important! Hmmm … that calculation looks wrong. I don’t really understand it, but I’ll change it anyway.

Application Configuration Can’t Be Controlled

VB supported? Maybe, maybe not. Analytics add-on pack? Maybe, maybe not.Third party optimizer? Maybe, probably not. Etc.

Good Data Goes Bad And Nothing Can Be Done About It

Just like every cell division results in some sort of degradation, be it a shortening of telomeres, an RNA transcription error, or the wrong number of chromosomes, every time a spreadsheet is copied or propagated, new errors are introduced. (That’s why 80% to 90% of spreadsheets have serious errors!)

Freshness is Fleeting

Like a loaf of bread, a static spreadsheet goes stale and gets moldy quite quickly.

Version Control is Impossible

The organization can define all the naming conventions it wants, but people are human and even if they try to follow the standard, they’ll screw up and the repository will degrade quickly. Plus, what happens when two people work on the sheet at the same time and upload a new version at about the same time. Whose is right?

The Sheet Is Not Even Guaranteed to Load

The MS add-in interface is poorly specified and that add-in could easily blow up the sheet, and if the user is unlucky, her installation.

In other words, SPREADSHEETS = FAIL. Is that clear enough?

Share This on Linked In

A Hitchhiker’s Guide to e-Procurement: Sectors

Mostly Harmless, Part XXIII

Previous Post

Enterprises generally fall into two categories: public or private. While the basic requirements for e-Procurement are unchanged whether the installation is intended for the public or private sector, each sector has its own standards for procurement and the system must support the standards, and quirks, of the sector in order to ensure adoption and success.

Since most of the posts to date have implicitly assumed that the application is going to be deployed in the private sector, this post will primarily discuss the public sector and some of the sector specific procurement processes that should have an impact on system selection.

In the public sector, most purchase orders for goods and / or services over a certain value are the result of a public competition. This means that once a requisition is approved for new products or services above a certain value, which can not be assigned to a standing offer or contract, the next step in the process is an RFX or e-Auction. As a result, the e-Procurement system will either need to contain RFX and / or e-Auction functionality, or integrate with such a system. The integration can be as simple as XML file export (of the details of the approved requisition that needs to go do bid) and import (of the details of the winning bidder and submitted pricing), but the workflow has to support the process and integration.

Also, as alluded to in the previous paragraph, new purchases below a certain threshold can be made against standing offers / contracts as long as the appropriate policies are followed. Some organizations will allow a buyer to use the standing contract of their choice (provided a certain dollar threshold is not exceeded in any calendar year), others will dictate round-robin selection in an effort to insure fair allocation of funds, and another group of organizations will use a hybrid policy and allow buyers to select the supplier from a set of preferred standing offers or offers that have not received their “fair” share of business.

Another area of increased complexity is approvals. In the private sector, the rules are usually cut and dry. For example, under 1K for approved products, the employee only, under 10K where all goods that can be bought on contract are bought on contract, the supervisor, under 100K, the VP, and the CPO’s approval is only required if the purchase is over 100K. In the public sector, there are approvals based on value, based on contract vs. standing offer vs. one-time buy, repeating payments, type of good or service (engineering will have to sign off on machine parts, HR on temp labor, etc.), MWBE percentages (as contracts over a certain value may be required to have MWBE components), environmental sign-offs (if the buy is for products or services with a significant environmental impact or products or services for which there are environmental standards), etc. Where it is quite uncommon for a requisition to require more than three approvals in the private sector, many requisitions can easily require six or more approvals in the public sector.

Payments are a little bit trickier too. Not only must they be processed according to standard terms and conditions, but they generally must be made on regular payment dates, and may be withheld indefinitely at any time due to orders from higher ups, which is common when governments have not approved their budgets for a year by a given deadline. Plus, many government departments are tax exempt (as it doesn’t make much sense for a government to tax itself) from local, state, and federal taxes, and the taxation rules need to be powerful and flexible.

Finally, the reporting requirements in the public sector are much more onerous than in the private sector. If the tool does not contain a modern analysis and reporting package that can meet all of the requirements of all of the departments and divisions and associated reporting requirements, it will need the ability to do a full export of all data required to produce those reports in a third party analysis or BI tool.

Next Post: Terminology

Share This on Linked In