Finding Your Way Around E-commerce

Introduction

Breaking into E-commerce can have you going in circles. That's not necessarily bad - you may not be sure where you are or where you're going, but there's a limit to how lost you can get. Even so, it's natural to want to get your bearings: E-commerce can be a dizzying ride. One report predicts that online sales in just one niche market will soon reach $3.5 billion. Another says that E-commerce will save businesses trillions in just a few years. Pretty exciting stuff, and everyone in the company wants a piece of it. But in the small print those same pundits suggest that the costs of building an E-commerce site top a million dollars, that most projects end up late, and that nobody really has a clue how to measure ROI. Would the best thing be to circle the wagons and start shooting?

As always, the way to get around the uncertainties is to have a map. Some maps are very precise, which is great when you can get one, but when trudging through uncharted territory its enough to know the major landmarks and where they lie in relation to each other. Any topographic information is a real plus.

The map in Figure 1 is a guide to the E-commerce space. Partitioned into three concentric circles, it can both prepare you for your trip and keep you from bumping into the bigger obstacles.

The Inner Circle

The Inner Circle (banish any thoughts of Dante) represents the basic infrastructure - both hardware and software - that's needed to get started in just about any form of E-commerce. You start walking counterclockwise from the place called IT/Networking. Even without specific web experience the knowledge already in this department will be the foundation for everything else. You'll want to cover all of the bases to get out of this circle: Note, by the way, that in this ring there are many opportunities for outsourcing.

• Routers: You already have routers for your network, but there are special models (and considerations) in web space.

• Firewalls: Depending on the kinds of web usage and applications, a router may not provide all the security you need. A firewall - technically a software product but typically running on its own small server -can offer greater security and flexibility.

• Servers: The riskiest thing about choosing a server is that you won't be able to get a reliable prediction for the load it's going to carry. What you can be sure of is that once an E-commerce site goes up, any little glitch is going to cause palpitations in the executive suite. So you need to look at fault-tolerant or redundant hardware solutions.

• Mail Servers: A mail server is software, and you probably have one already for corporate e-mail. But E-commerce can inflate both the magnitude of e-mail and the functionality needed from the server, so it's not safe to assume that the solution you have will stand up to the job. Long-term performance considerations may dictate that this live on its own server, but that's a decision that may not be required at the start. Fax serving may also be a consideration, especially in a B2B web site.

• Storage Systems: Again, the requirements of reliability and continual uptime may cause you to investigate different models from the ones you're used to.

• Web Servers: This is the software that serves the web content to the user's browser. Unfortunately, while the major players all have about the same functionality, this will prove to be a religious decision (Netscape vs. Microsoft, open vs. proprietary, …) that is going to involve IT staff, programmers, and web designers. These battles actually started up above when you picked the server hardware, didn't they? Note, by the way, that you'll see vendor offerings for "Commerce Servers." These are webservers enhanced with support for certain commerce activities. Use them out of the box or as toolkits for custom solutions, especially where a one-vendor solution makes sense.

• Databases: Another area where there is certainly core competency in the IT/Networking department, databases can present special challenges in a few ways. While the OLTP model is important, many websites require that information be sliced, diced and extracted in ways that don't easily fit the relational model. Also, the rate at which they are hit can exceed what you've seen in traditional IT applications.

• Backup: Even something as basic as backup (and recovery!) has its own web spin. The well known problems at E-Bay highlight the need for redundancy at all levels, but unlike more traditional IT approaches, there is no overnight window for periodic backup, and delays of hours in reloading a database won't be tolerated. ("Dammit, Scottie, I need warp drive now!" "Aye, Captain, I'll do me best.") Meeting business needs here requires a combination of software, hardware and network approaches.

The Magic Ring

What really puts the squeeze on is the tools and applications in this layer. A few years ago they were thought of as goals in themselves, but they re quickly becoming de rigeur foundational elements that the business side will take for granted in designing their web strategies. Unlike the Inner Circle, there is no one clear path here in the Magic Ring. Any particular site strategy might call for a different subset of these functions, and might require them to be implemented in any order.

• Security and Certification: Beyond security at the router and firewall level, here is where one gets into such considerations as encrypting credit card and other personal data, digital signatures and, in general, proving to customers on a real-time basis that they're dealing with whom they think they are.

• Development: There are a wide range of tools, from simple HTML editors to complete web page development environments, from simple forms to complicated Java applets or Active-X components, that can come into play during the construction of a web site. Even when development is done outside of IT, the IT department will get involved because some of these approaches have server and database implications. And, as with the basic server decisions, this is an area where religious wars can sprout at the slightest provocation.

• Testing: At web speed there's never time to test: Put it up first, and then find the problems; if performance is slow, throw on another server. Those who are uncomfortable with this approach, or who want to have an approach more organized than the Hundred Monkeys Testing Methodology, will want to invest in tools that test content accessibility, basic functionality, and behavior under controlled load. For the rest, there's stress testing - except it may be that it's IT and not the web site that gets graded.

• Content Management: If managing network configuration files and software versions is getting routine, Content Management may prove a bit more interesting. At the low end there's the problem of size. A web site can throw up hundreds of pages per week on an ongoing basis, but those doing the throwing may be writers or marketing types who've never thought about the problems of cleaning out dead files and tracing the chain reaction caused by small changes. At the highest end, companies whose business is content (ranging from publishers to B2B vendors to catalog sites) may be interested in tighter control of style and format, and in easy repurposing. There are high price-tag applications that do content management on a corporate basis. These may have enterprise-wide benefits; they're sure to have enterprise-wide consequences.

• Ad Serving: A critical technology for almost any web site. Even if your site is not serving ads, one needs to pay attention to this technology. We're starting to see products that support your advertising on other sites. (Even though your ad agencies will support your advertising efforts, there may be some technology issues involved with bringing the reports and the control to the right desktops.)

• Traffic Reporting: You might not believe how important this is. In theory, the basic reports could be developed from the web server logs by an intern, but you'll find that many different departments want different kinds of packaging and details. Traffic reports are their clue to what works and what doesn't, and the faster they can find out about a page that's too boring for people to read, or one that's a grabber, the faster one can make changes that have a direct or indirect effect on revenues. This may be the one piece of web technology that everyone on the business side will use - and scream about when the reports are late or inconsistent.

• Data Mining: It isn't clear what the eventual name for this area will be, or exactly what its functions will be. What seems to be evolving is a data warehouse approach - that is, quick inquiries into a wealth of information. It's built on traffic logs, advertising reports, and some kind of real-time analysis that captures information about user behavior that the other sources miss. If someone in marketing wants to ask how many people saw the widget ad on the web newspages exactly three times before they clicked on it, or how many times people will search for something before they decide it can't be found, this is the technology you'll need. It's an area that is intimately associated with Personalization.

• Personalization: This is in part the flip side of Data Mining. Given detailed knowledge about how individuals (or groups with certain behavioral characteristics) behave, how can we personalize their experience? More than a MyPage approach, personalization refers to serving ads or reordering content in real time to match these behaviors and get people to stay on the site, click on ads, or buy. When a site requires registration or uses technology to differentially identify anonymous visitors, the potential for using mined data increases many fold.

• Messaging: How many ways can E-commerce create a message? Automated responses, targeted newsletters, fax blitz, groupware, IP telephony, paging, and who knows what else. Communication is still the foundation of a business, and the Web only multiplies it.

• Help Desk: A necessity for any business over the web. Naturally, the people at the desk will need tools for accessing the various databases used by the website, including the data about user behavior and purchasing.

• Relationship Management: Part Help Desk, part Personalization, the term is used by some to mean the kind of customization of user experience that the web makes possible, and by others a natural web (or intranet) extension of traditional tools for managing more direct contacts with customers and vendors.

• Extranet: An extranet can be an important way to create a secure, closed network between you and your business partners.

• Search: At the very least you'll need a search function to help people find things on the site. There are various options that differ on usability, utility, and the way that they reflect your content. Some sites will supplement their content with links to related material that's found by searching the web.

The Beltway

Out here in the fast lane are the applications and functions that define the web site. There are new tourist attractions showing up every day. Since no attempt at a complete list will be complete for very long, we've only shown a sampling that covers both villages and the capitol cities.

• Survey: A web site is a great way to gather information about things that interest you: Just post a survey and get a few (hundred) thousand responses. Posting the results on the website for users to see generally adds more page and ad views.

• Community Creation: Many web sites support forums (bulletin boards as they used to be called) and chat. A good way to pull in visitors and sell ads -- especially if your site won't be hurt when people exercise their free speech. The technology here can be very flat or allow users to create animated characters to represent them in chat space.

• Catalogs and Directories: Although there aren't firm distinctions here, a catalog is generally a product list that leads to a purchase, where a directory is a searchable database that leads to information or links. In B2B applications a catalog may be hosted by the vendor, by a third party that brings together a number of vendor catalogs, or by the buyer.

• Storefronts: If you're selling to consumers, the storefront is where you do it. Storefront software will probably include support for building a product catalog and handling credit card transactions.

• Auction: It isn't all that long ago that eBay burst upon the scene, and already there's packaged software that lets you put an auction on your own site.

• Marketplace: A marketplace brings buyers and sellers together. For a company that has a presence in both camps this could be an option for new revenue opportunities.

• Payments and Billing: This really is the point of it all, isn't it? In a consumer environment it's a relatively simple matter of accepting credit card numbers and passing them on to a processor, and possibly interfacing with an e-money or wallet vendor. The B2B world is quite different, with the need to access back-end financial systems and the requirement to generate reports that can be handled automatically by your partners' back-end systems.

• Human Resources: With rsum gardens like MONSTER Board and HotJobs, there is every incentive for a company to do more recruiting on the web. Placing ads on websites or in newsgroups is just the tip of the iceberg. The more that resumes can be accepted from e-mail or forms submissions and processed automatically, the faster they can be in managers' hands.

• Industry or Market-Specific Protocols: Global, business-to-business E-commerce demands that information flow easily between companies and between applications. And that implies standards, created by impartial groups supported by all vendors. And that could happen? But until then we get to choose between proprietary, vendor specific approaches and vendors or partners supporting one of a number of emerging standards.

So there's a lot of territory to cover when someone says "We've got to do E-commerce - everyone else is!" But a careful approach, starting with corporate and IT strategies rather than jumping into a particular application or technology and hoping that the next 10 initiatives will plug right in, always has been, and always will be, the best chance for success. And you'll probably discover that E-commerce isn't really so bad - these days it's found in all the best circles.