Daily Archives: May 25, 2010

More Reasons the Cloud is Not a Fluffy Magic Box

Soon after I told you that the cloud is not a fluffy magic box, I found this great post over on an Information Week blog on 3 things that could kill the cloud which points out some more sobering realities of the cloud, which is just really an abstraction of the multi-tenant SaaS model where one provider provides the software and another provides the infrastructure the software runs on. The article has some good points that should be taken into account before you decide that the cloud is the answer. (Sometimes it is, sometimes it isn’t.)

  1. Scalability is not Unlimited

    First of all, at any point in time, the infrastructure provider has a limited amount of hardware and bandwith available. When that is reached, you’re out of scalability until the provider ramps up. Furthermore, even if the provider ramps up, there’s still a practical limit dictated by the software. Most databases start to fail miserably when you get to the Terrabyte range. Most analytics applications fail miserably when you ask them to process millions of records in real time. Etc.

  2. Security is not Absolute

    The cloud does not inherently provide more security as some vendors would have you believe. In fact, it might even provide less. In reality, the security of any platform comes down to the knowledge and vigilance of the provider’s people and how well they are at identifying potential holes, locking them down, and keeping up with patches. If the software vendor assumes a certain port will be locked down and the infrastructure provider leaves it open or if the hardware vendors assumes the software vendor will patch core applications and vice versa, security is weakened.

  3. Prices can be Higher

    While up front prices are quite cheap as you’re primarily paying for energy costs (to run and cool the CPUs) and bandwidth, and while the Cloud will be cheaper for small-scale applications, the reality is that for large scale, high-bandwidth, applications, the total costs can be more expensive than running your own data centre as most providers don’t yet have the scale and expertise to beat in-house costs. You have to do the analysis.

  4. Your application can disappear in a puff of smoke.

    Thanks to the Patriot act, if a drug dealer happens to be using the same multi-tenant provider, in the US the FBI can sweep in and seize *every* server in the data centre, regardless of what else is on the servers, shutting down the entire operation of the infrastructure provider for an unspecified time — like they did to Core IP Networks in April.

Share This on Linked In

Sometimes Old-School Works Just Fine: An EC Sourcing RFP Case Study

It’s still a buyer’s market. Many suppliers are desperate for business, supply (capability) still exceeds demand in many markets, and even though prices are starting to rise in some markets with expectation of recovery, they haven’t risen much yet. According to many strategic sourcing professionals, it’s the perfect market for an (e-)Auction because suppliers will compete for your business. And while that may true, it does not guarantee that you’ll get the best result.

An e-Auction carries a number of risks. The result can be higher prices than a traditional negotiation. For example, if the auction was limited to a small number of suppliers, who are in contact, they may collude to keep prices high — or they may all adopt a strategy of delayed bids and minimum bid decrements which could result in higher prices. The result could be unsustainably low prices. A supplier, desperate to win business, might hope to make up losses in future volume, bid a razor thin margin, and then risk bankruptcy when its costs rise. At this point, the only choice for the organization would be to accept higher prices (through surcharges) or risk an interruption while a search for a new supplier was conducted. And the result could be strained supply base relations. A poorly conducted event can instill animosity in winners and losers alike, which would result in poor service from the winners and lack of response in future bid requests from the losers.

As a result, sometimes the best approach is an old-fashioned multi-round RFX with feedback between each bid, as it was for a certain mid-size apparel retailer, who we’ll call Apparel-For-You, who was new to e-Sourcing and just wanted a way to streamline their ocean freight bidding efforts (for their 25M ocean freight category) and communicate with suppliers in a consultive way. Specifically, Apparel-For-You, not realizing the significant savings opportunity before them, had the following goals in their search for an e-Sourcing solution:

  1. Understand Supplier Willingness to Bid on a Per Lane Basis

    Historically, Apparel-For-You had been surprised a number of times not only with respect to bids that came in, but with respect to lanes carriers were willing to bid on individually

  2. Reduce Analysis and Reporting Time

    Apparel-For-You’s supply base, which provided them with over 2,200 individual SKUs, was spread across 30 ports of origin, 4 major ports of destination, 9 carriers, and 4 container types — which equals 4,320 bids to be collected and analyzed before the 120 lanes can be divided among the carriers. While certainly not impossible to do by hand, that’s still 10 (9 carrier plus 1 integrated) fairly large spreadsheets to manipulate and analyze in a time consuming and error-prone manner.

  3. Communicate with Suppliers in a Consultative and Regular Fashion

    Without a dedicated sourcing tool, it’s difficult for all team members to know when a carrier was last contacted and what was discussed. The ball could be dropped, and this could lead to a damaged relationship. Given the importance of relationships in Apparel-For-You’s supply chain, as apparel has a short product life-cycle, this is something Apparel-For-You wanted to avoid.

Given these requirements, and Apparel-For-You’s lack of e-Sourcing sophistication, EC Sourcing recommended that Apparel-For-You use a multi-round RFX, starting with an RFI to find out which carrier was interested in which lanes, with analysis and feedback between each round. The carriers were all informed up-front of the new process, and Apparel-For-You consistently followed-through after each round.

Using the built-in templates, Apparel-For-You was able to easily create an RFI that allowed it to create the right pricing matrix for each supplier as well as clarify important T&C’s with each. The process of collecting bids from carriers, who were used to Excel, was simplified by way of Excel integration. This integration also simplified the amalgamation of the bids into a single matrix for analysis purposes, as the integration was automated and free from human error.

Using built-in reports and advanced analysis models provided by EC Sourcing, Apparel-For-You was able to quickly analyze the bids after each round and provide the supplier with feedback on their relative ranking, which included how much they’d have to lower their bids to improve their rank and take the #1 spot. Using this information, the carriers were able to adjust their bids accordingly and focus on the lanes they could perform the best on with respect to the buyer’s needs.

In the end, Apparel-For-You not only accomplished their goals of

  1. Understanding Supplier Willingness to Bid on a Per Lane Basis

    as this information was known before the first bid was collected

  2. Reducing Analysis and Reporting Time

    as the project time-frame was reduced by 35%

  3. Communicating with Suppliers in a Consultative and Regular Fashion

    as they were able to inform the carriers of their rank and potential awards promptly after each bid, and track when the last discussion took place

but Apparel-For-You also reduced their costs by 19%.

This just goes to show that, sometimes, old school works just fine. The full case study is available in PDF form.

Share This on Linked In