Optimisation Algorithms

Last week’s Economist (15 September) has a special briefing on algorithms. The 3 page article covers a range of applications for algorithms:

  1. Making sense of unstructured data (Autonomy)
  2. Routing aircraft in an optimal way (UPS)
  3. Routing calls to call centers (Convergys)
  4. Making networks more resilient to data disruptions (BT)
  5. Calculating optimal location of goods on shelves (Tesco)
  6. Looking out for potential fraudulent activity (ClearCommerce)
  7. Finding the most relevant results for an internet search (MSN, Yahoo, Google)

Algorithms are also prevalent in optimisation routines which help buyers decide the best way to award contracts to competing suppliers. Yet not only did The Economist fail to quote Michael Lamoureux, they didn’t even mention procurement. To be fair they did mention supply chain, but supply chain attracted dramatically less words than even call centers did.

To me this means one of two things:

  1. Procurement is an uninteresting dead end, compared to sales, marketing, finance and technology
  2. Procurement is still a nascent, poorly understood discipline that has yet to make its mark on the general consciousness.

Obviously I’m in this game because I believe option 2 is the case ….

Procuri: Another One Bites The Dust

Jason over at Spend Matters has the big story on Ariba buying Procuri.

I remember when Ariba bought (sorry, merged with) Freemarkets. My initial reaction at the time was that a major competitor (Freemarkets) was now going to be bigger and stronger than ever. But in reality everything that Freemarkets had to offer disappeared, as if Freemarkets had never existed. Ariba continued to sell software as if nothing had happened and the old Freemarkets products and services ceased to exist.

I wonder how far Ariba will go in obliterating all memory of Procuri. How much of Procuri’s TotalSource will find its way into Ariba’s toolset? My guess is not much. Procuri’s software is reasonable but I doubt any of their kit is streets ahead of Ariba, who also have software covering roughly the same space. Will Procuri be re-branded as a mini-Ariba and kept as a separate entity, or will it just disappear into a puff of marketing?

Whatever, these are exciting times. Thanks to Jason for spreading the news!!

Sun’s e-Auctions

Great article over at Purchasing.com http://www.purchasing.com/article/CA6474837.html?industryid=23508. It’s an interview with Kurt Doelling, Sun’s Vice President of Supplier Management in which he talks about Sun’s ongoing use of reverse auctions (or Dynamic Bidding Events, DBE’s, in Sun’s Parlance).

Apparently, of Sun’s $4bn annual spend with suppliers, $2bn or so is awarded via DBE, expected to rise to $2.7bn this year. That is a very impressive statistic and means they must be doing a lot right as far as e-auctions are concerned:

Some commentary on Kurt’s specific points:

  1. Reverse Auctions are based on Total Cost of Ownership rather than just unit price, with suppliers being score-carded on attributes such as availability, technical support and quality. This is great.
  2. Don’t invite rogue suppliers just to drive the price down. This is absolutely paramount. Inviting low-priced suppliers that you have no intention of doing business with purely to drive the price down is tantamount to shill bidding (which is fraudulent). Certainly it damages the credibility of your sourcing process.
  3. Sun only invites known suppliers to their reverse auctions. Whilst this facilitates the previous two points, it is interesting to learn that they would not include new potential suppliers in a reverse auction. This suggests that Sun already knows, and does business with, the vast majority of their potential supply base.
  4. Reverse Auctions are a way of transferring margin from suppliers to buyers. Blunt and honest. Presumably this is an important driver for all buyers.
  5. Award shares of business. It’s common sense, but good to hear a practitioner advocating sourcing from multiple suppliers rather then just one.
  6. Invite 3 suppliers to an auction. Although this seems to deliver good enough results for Sun this seems a bit on the low side to me.
  7. Only show suppliers their rank rather than best bid. When I was auctioned, being shown rank only was a demotivator to bid. Typically, auctions I have seen work better with best price being shown, assuming you can generate enough competition in the reverse auction.

E-Auctions (hunh). What are they good for?

In the immortal words of Melbourne tough guy Johnny Morrison: “[R]eputations are based on a hundred facts – and a thousand fairytales. ” (*)

There are plenty of facts and fairytales surrounding e-auctions and e-sourcing. One that grates most with me is that e-auctions are about cutting costs and e-sourcing is about improving processes.

True, e-auctions (with all my usual caveats about them being well managed etc) will deliver a lower price than the same sourcing project done without an auction. And true, companies like TradingPartners do tend to sell e-auctions as ways of saving money, so to some degree we only have ourselves to blame.

But here are 2 subtle facts about e-auctions that often get missed amongst all the fairytales:

  1. In order to run an effective e-auction you need to do your homework up front and be crystal clear about specifications, award processes, requirements, service levels etc etc. In other words you have to run a robust sourcing process.
  2. An e-auction can’t magically deliver savings on its own. In a rising market you may well end up paying more than you otherwise would. However, an e-auction is your best bet to get the most competitive market price and will generally get you a better result than what is possible by other means. Assuming you plan, manage and run it well.

Whether the extra effort in running a robust sourcing process and planning, managing and running an e-auction is worth the $$$ benefits is a different question, of course.

The Buxton theory of cost reduction holds that falling house prices and/or falling markets put cost cutting onto the CEO’s radar, so I fully expect e-auctions to become used even more widely over the next couple of years. On one level this is good news for an e-auction provider like TradingPartners. But is also has a downside because it reaffirms in people’s minds the fairytale that e-auctions need only be about saving money.

(*) Quoted by Mark “Chopper” Read in his book Chopper. The full advice from Johnny Morrison is: “Mark, reputations are based on a hundred facts – and a thousand fairytales. A horse is only as good as its last race. So get in there, rip his head off and piss down his neck.” The second part is probably not so directly relevant but I thought I’d post it for entertainment value as much as anything else.

I was auctioned part 2

In a previous post, https://alanbuxton.wordpress.com/2007/09/05/i-was-auctioned-part-1/ I gave some details about a reverse e-auction I took part in as a supplier and explained that we held back in our bidding. Here are some of the reasons why:

  1. Lack of clarity of what was being sourced. The documentation accompanying the e-auction was only about 1 page long. So even if though we had our own interpretation of what we were bidding on, how could we be sure that we were bidding like for like against other bidders, or even whether the first placed bidder was a genuine bidder or a shill.
  2. Lack of competitor information during the e-auction. This particular format only showed us whether we were first, second or worse.
  3. Lack of professional e-auction management. The e-auction manager did not give us any reasons or encouragement to bid during the event
  4. Lack of clarity over the award process. When would it be awarded? Would it be awarded at all? Would it be awarded to the supplier with the lowest price?

With all this in mind it seemed to me that we had two choices:
A. Take a very narrow definition of what the buyer was sourcing, bid very aggressively based on that interpretation and then make money through charging for add on services.
B. Take the negotiation offline if possible. Place some bids during the e-auction but not bid too aggressively and then try to talk to the buyer after the event.

We chose option B. We weren’t awarded the business so to some degree this is an academic debate. However I am sure that many suppliers who take part in e-auctions ask themselves exactly the same questions. If we had been encouraged to bid more aggressively during the e-auction, who knows, maybe more competition would have been generated and maybe the buyer would have achieved better prices.

The moral of the story is for buyers to approach their e-auctions in a professional way, to ensure they are completely clear with suppliers about where the e-auction fits into the overall sourcing process, and to make it worth the while of suppliers to bid as aggresively as possible.

Buxton’s Hierarchy of Technology Needs

This started as a bit of fun but it’s turned out to be a surprisingly useful model for me. With apologies to Maslow’s hierarchy of needs this is Buxton’s hierarchy of technology needs.

Maslow’s hierarchy, from top to bottom is:

  1. Self-actualisation (e.g. morality, spontaneity, creativity)
  2. Esteem (e.g. achievement, confidence)
  3. Love/Belonging (e.g. family, friendship)
  4. Safety
  5. Phyisology (e.g. food, water, shelter)

The point being that when you are being threatened by a bear you are not going to be particularly concerned about writing poetry.

Buxton’s Hierarchy of Technology Needs, then, from top to bottom goes as follows:

  1. Game-Changing (e.g. development of a new generation of products)
  2. Esteem (e.g. delivering better processes, directly supporting the generation of revenue)
  3. Social (e.g. having core applications like everyone else does- Internet, Excel etc)
  4. Safety (e.g. having a reliable PC that doesn’t crash)
  5. Communication (e.g. just having a phone or fax that will allow communication)

The point for technology leaders being that if people’s PCs keep crashing then the new SAP system you’ve just implemented is not going to be very interesting.

Some technology leaders are happy with working at just the bottom 2 or 3 levels. Some organisations consider technology is only good for the bottom 2 or 3 levels. At these levels technology is about providing “fixes”.

However as the technology leader you have a great opportunity to invent solutions for problems that have yet to be identified. (Invention, after all, is the mother of necessity – see this earlier post). When you do, just don’t forget the importance of the things at the bottom of the hierarchy.

On Google, Banks and Supplier Networks

The Economist, Sept 1st, contrasts Google’s development with other tech companies

Google is often compared to Microsoft; but its evolution is actually closer to the banking industry. Just as financial institutions grew to become repositories of people’s money, and thus guardians of private information about their finances, Google is now turning into a custodian of a far wider and more intimate range of information about individuals.

When a company sells you software for you to install on your PC (e.g. Microsoft office) that transaction is pretty much the end of your relationship. However each time you go onto Google you are sharing your search habits and some level of browsing habits. If you are using gmail you are sharing a hell of a lot more. As you pile all this information up the value to Google of the information increases exponentially.

This is potentially the same with all Softare as a Service offerings. In the procurement world the best parallel would be supplier networks. Today, supplier networks are generally used to facilitate the sending of transactions backwards and forwards – like a form of EDI that happens to use XML and the Internet. A more Google-like approach to supplier networks would capture buyer and supplier interactions in such a way as to then leverage this information to be able to match buyers with suppliers more effectively.

Some more stats on Google’s meteoric rise, again lifted from The Economist:

Less than 10 years old
Market cap $160bn
Expecting Revenue this year of $16bn
Expecting profit this year of $4.3bn
13,786 employees in June 2007