- Introduction
- Overview
- Fear and Greed
- Rules of the Game
- Summary
Rules of the Game
To be successful, web development groups need to learn the rules required for surviving and thriving in the new environment. Each rule illustrates an essential reality in the today's transformation of business processes as they incorporate the Internet into everyday work. Because these rules govern the basic precepts of the Internet, familiarization with these rules will be essential to an understanding of the discipline of content management.
Rule 1: It's the Assets, Stupid!
Founded in 1927, W.W. Grainger is one of the largest business-to-business (B2B) players. It sells 220,000 maintenance, repair, and operating (MRO) items through 500 stores to 1.5 million customers. Its annual sales run at $4.7 billion. Back in 1995, the Grainger management team seeded the web operation with an initial infusion of $5 million. Today 150 people work in the online operation, and in 2000 it expected to do $200 million in business over the web.
The key to Grainger's e-commerce franchise goes far beyond digitizing the content contained in its red phone booksized paper catalog that lets businesses browse and order items ranging from axes to rubber boots, to voltmeters. Digitized catalog items, consisting of images, textual descriptions, and pricing information are the most highly visible asset. But the value goes much deeper. The Grainger site hosts information areas that range from hurricane preparedness tips to proper care of power tools. These help areas put a friendly face on the staid corporate look, and they build traffic and goodwill with customers. Behind the item information lie item availability information, search keys, and indexing information. They power the search and indexing capability that is a key differentiation for the online catalog, as compared with the traditional paper catalog. Finally, application code drives the entire web experience; it defines the business logic that governs ordering, shipping, and customer account management. Running a web operation of the magnitude of the Grainger site requires a wide range of assets, many of which aren't directly visible but nonetheless are indispensable components of the operation.
It is the web site assets, taken together, that power the Grainger web property. They are irreplaceable. The business cannot function without them. At the same time, the assets change. Some assets change frequently. For example, a hurricane preparedness site and a blizzard preparedness site are seasonal. Some assets change less frequently. For example, new corporate privacy policy introduces new logic to the ordering sequence, and that logic remains intact for years. Above all, the constant evolution captures the collective learning of the entire organization. The web assets embody this learning. The web and associated server give the assets life, which, in turn, express the organization's collected wisdom in an active form.
Rule 2: Experiment, Iterate, Grow
A corporate research and development center is an unlikely location for cutting-edge forays into novel business models, but that's exactly where innovation began in a multibillion dollar global transportation and logistics company in 1994. A researcher inadvertently steered very different two worlds onto a collision course. One of those worlds was the public Internet, the outpost of freewheeling renegades operating at lightning speed at the boundaries of the law, or sometimes beyond. The other world was the corporate back office, a place traditionally dominated by big-iron mainframe computers and buttoned-down security procedures. After uniting the two worlds, vast possibilities unfolded that companies routinely exploit today.
So what was the historic act? The researcher rigged a web server outside the corporate firewall to query package shipment status and to present the result on a web page. Armed with a shipper's package identifier number, a sender and recipient can instantly determine the exact whereabouts of the package along its journey from sender to recipient. It was a revolutionary idea. If a Hollywood screenwriter had scripted the event, the protagonist would have grasped thick cables in each hand and brought them together in an explosion of sparks.
A quiet revolution followed when the researcher built an integrated system to present detailed information about package shipment status hosted on a corporate mainframe. Connecting the corporate back office to a newfangled web server attached to the public Internet opened the door to countless innovations to follow.
From the moment that the two worlds fused, the ensuing events moved in a way that has been repeated in many businesses and organizations, in many different forms. A new force emerges, which compels organizations to embrace the Internet as a medium for disseminating information, realizing commerce opportunities, and bringing new opportunities to fruition. Just as innovations of the past, such as steel making, the railroad, or the telephone, opened the doors for new arrangements of people and resources and new business opportunities, the Internet causes gradual yet unmistakable realignment of operations.
This is the irony. The researcher was merely goofing around. It is unlikely that he fully grasped the immense repercussions of his actions. Instead, he modestly set about to demonstrate the feasibility of public access to up-to-the-minute internal logistics data over the Internet. The technology allowed it. The opportunity was there. He seized it.
Several years later, the business model pioneered by the transportation and logistics company would be given a name: online customer self-service. In this model, a customer uses the Internet to access a web property that allows him to track the status of an in-transit package whenever he chooses. The corporation reaps benefits because the new service is a low-cost and efficient way to improve the responsiveness of its customer service operation.
This example is instructive. The researcher's actions brought dramatic changes. This example illustrates an important heritage that it is vital for web development teams to remember. Big changes sometimes begin with goofing around. An artist experiments with a new juxtaposition of graphic elements. A developer tries out a different twist on a Flash-enabled page. A SQL developer wants to try a different variation about presenting the results from a query. Big changes arise from small innovations. Each contributor needs a place to try out these changes without needing to endanger the entire organization's effort by having everyone share a common staging server.
Iterative development, proven to be immensely valuable to developers, is even more valuable to web developers. Innovation springs from experimentation and iteration. That is a fundamental truth of web development.
Rule 3: Respond to Customers Quickly and Frequently, or Lose Them!
Move ahead to 1996. Inspired by the early work of the playful researcher, the global transportation and logistics company builds package tracking into its public web site. The distinctive feature gains popularity among the customer base. The company touts the ability to track packages as a unique service offering. The package-tracking site has become part of company's public persona.
Meanwhile, the success of the web site brews another problem. The staff to support the web site has expanded to include more than a dozen people in three departments. Marketing scopes the requirements for the site and engages with outside vendors to implement the look and feel. Information technology incorporates changes, integrates with the mainframe connectivity code, and takes responsibility for the day-to-day management of the web assets. The operations group moves the modified assets onto the multiple production servers, and they maintain 24x7 uptime.
Changes pour into the site to respond to customer feedback and new marketing requirements. Uptime must be maintained throughout. Efforts of dozens of contributors from three different organizations must be coordinated. Incidents that require immediate attention by the operations staff occur more frequently. Handling emergency fixes to the external web site becomes a daily routine. Everyone realizes that there needs to be a better way.
This is a recurring theme of web development. Early experimentation yields promising results, which encourages larger and more formal business initiatives. Initial efforts are usually modest, involving small staff and humble resource expenditure. Follow-up efforts involve more people and bigger budgets. Time pressures swell. The stakes escalate. The coordination difficulty magnifies, and the cost of failure is greater. Management pays more attention, compounding the level of stress felt by everyone.
Operators in the Internet space push the envelope of what has been done before, beyond commonly standard practice. The pace of change is relentless. The practice of content management embraces the need for rapid turnaround because inferior efforts are discovered and punished, often in real-time. The Internet encourages side-by-side comparisons of web properties.
Let's see how this works in a simple example. Suppose you have an inkjet printer, and you need to purchase a replacement ink cartridge. You decide to purchase over the Internet. You have many online vendors to choose from. Let's suppose that you haven't settled on a vendor and that you have come to rely on for price, delivery, and ease of online ordering. You probably do what many others have done, which is to simultaneously browse multiple online stores. You type the brand of your printer. You probably don't know the manufacturer part number. You obviously don't know which vendor will have your item in stock.
You decide to shop two sites side by side. You shop one site in one window, and, while you wait, you open a second window to visit a second site. Side-by-side online shopping is unlike shopping in the brick-and-mortar world. Store operators in the physical world understand clearly that each step that you take into their store increases the likelihood that you'll complete your purchase there. If you ask a clerk for assistance and your request is ignored, you might make a quick getaway to a competing store.
In the online scenario, if the search mechanism on one site is clunky, your attention drifts to the other site. If the application server behind one site is slow to render the page, if the servers are overloaded, or if the flavor of JavaScript used on the page isn't compatible with your browser, you'll be favoring the other site in no time at all. If both sites display the correct ink cartridge but one site cannot tell you whether the item is in stock, the likelihood of completing your order with that vendor drops significantly. If one web site cannot tell you how many days before you'll receive your shipment in the mail, or if the unit price looks good but the shipping cost totals out to a higher final cost, your business evaporates. A recent study by Andersen Consulting suggests that nearly 90 percent of online shoppers have abandoned a shopping basket at least once during a shopping season. Another survey by Boston Consulting Group and Shop.org reveals that 65 percent of online shoppers abandon their filled shopping basket without purchasing.
Not every consumer engages in side-by-side shopping. But its availability points out the fact that web properties must always be vigilant to rapid defection of its customer base. Savvy web operators avoid this defection by rapidly updating their site's functionality. What was not fully comprehended yesterday becomes plainly obvious when customer feedback pours in, direct or otherwise. This type of rapid feedback has serious implications for the already overloaded web development staff. Requests for change drench the staff. Changes that can be completed quickly must be completed quickly. Changes must be tested, reviewed, approved, and moved into production. This must be done correctly, precisely, and with carefully orchestrated actions. Wasted motion must be avoided. Efficiency is everything.
Rule 4: Enable the Masses!
In 1996, the South of Market Area of San Francisco gestated a significant new web property. It was the home of a groundbreaking web publication, which produced online content combining the talents of writers, artists, editors, and developers to create a web-based virtual magazine. Unlike traditional publishing, the Internet and web technologies distribute content instantaneously to readership around the world. In this case, too, experimentation was rampant. Different graphic layouts, different navigation schemes, different approaches to integrating search, and different editorial styles were all used. For example, the three-area page layout style commonly employed today evolved during that era of experimentation and consisted of a header, a left-side navigation area, surrounding a main content area.
In the next stage, traffic to the property boomed. Scores of talented contributors occupied an upper floor of a renovated brick factory building. Computers, monitors, tables, and office chairs crammed into a bustling open-air working space. Editors, artists, and developers assembled edgy pieces for an avant-garde readership. Building repeat traffic to the site requires content that changed frequently and involves breaking away from the long editorial cycles familiar to magazine publishing. Content-wise it was a magazine, but its publication cycle made it unlike anything that preceded it. Virtual "issues" were published daily, sometimes hourly.
The contributors gleefully embraced the rapid pace of change. But the breakneck publication cycle slowly spun out of control. Work was overwritten. Changes were lost. Mistakes once corrected reappeared.
A Draconian mandate established a stopgap solution. A production manager decreed that there should be seven prescribed folders on a designated file server. Each folder would contain work products at a particular point in the development process. When work was complete at one stage, the intermediate work products would be dragged to the folder representing the next stage. Nominally, a project was to take seven days to move through all the stages. From the folder representing the final stage, the content would go to a staging server, where final quality assurance and approval would take place. If approved, the content would be copied into production on multiple redundant web servers. In the ideal scenario, each project would move methodically from folder to folder, projects wouldn't depend on each other, and projects wouldn't move backward. Frustration set in when the editorial realities intruded and files needed to be handled out of sequence.
Soon it was obvious that there was a need for a better way to handle concurrent development by scores of contributors, although the folder solution still represents a useful strategy. The creatives gravitated toward the Macintosh, whose operating system popularized the paradigm of dragging files between folders. The developers edited text files directly on Unix workstations. The seven-folder solution was merely a stopgap until a better solution could be devised. Anything is better than allowing uncontrolled change.
This is another recurring theme of web development. The simultaneous contributions from scores of developers must be managed in a fast-paced environment.
Rule 5: Make it Manageable and Reproducible
Barron looks at his watch. It is going to be tight. Although it is late at night on the East Coast, in just a few hours it will be morning at the London offices, and employees will click open the daily online company newsletter to read the breaking story about the restructuring. But it isn't the news story that worries Barron.
He has survived many restructurings in his career. In fact, Barron takes comfort in realizing that his business unit's express charter is to make sure that business-critical web content shows up correctly, at the right time, with enough bandwidth to supply the demand. His unit has put together a pretty efficient operation, he reflects proudly.
The CEO recorded a company-wide announcement earlier that evening in the corporate headquarters in Manhattan. The address has gone longer than was originally planned, which is the source of Barron's concern. Because of the longer duration, and because of the accompanying web assets to be placed on the intranet web servers, the network performance experts expect a higher than expected load on the multiple streaming video servers placed around the globe. Standard procedure is to use an automated content deployment system to preposition media files on servers strategically placed at key locations in the firm's global network. In order to handle the higher load without choking the entire network, Barron's group has been summoned to assist.
It is only because of the deployment infrastructure that Barron's group has built that it can respond by deploying the CEO's address and the associated web assets on additional servers. Barron can hardly believe that all of this has taken place the same day. All morning information technology staffers at key locations around the world bring new content server machines online. By the time testing of the new capacity is completed, only a few hours remain to specify the additional machines to receive the deployment.
Barron is proud of the decentralized administration. A crew in each geographic region has the ability to add the new machines to the local deployment tables. When each group finishes, they indicate their readiness. The first call comes in from the Hong Kong office. Barron isn't surprised; it is the middle of their workday. When the final call comes from the London crew, they are ready to proceed.
Barron wants to verify that the new capacity has been brought online.
"Let's edit the homepage for the corporate online newsletter," he suggests to Sarah. Sarah is the lead designer for the online newsletter. Her eyes scan the page for something innocuous to change. Her eyes stop at the text "Jan 10," which she deftly edits to "10-Jan".
"Let's push this change out now." Sarah suggests. Barron agrees. Both of them know what will happen next. Because Sarah designated this change a high-priority change, the modified homepage is immediately copied to Intranet servers around the world. Within a minute, Sarah's pager beeps. The message confirms that the modified page has been successfully deployed to all servers, including the new ones that have been brought online that evening.
They are nearly done. The servers are operational, and the content has been fully approved before the close of the business that day.
Sarah directs the content-management system to schedule a deployment of the CEO's video announcement to the intranet web servers to coincide with 7 A.M. London time. She double-checks that her pager will be notified if any problems occur. Because the corporate communications group wants the announcement to be released no earlier than 7 A.M. but no later than 8 A.M., that will give an extra hour if something goes wrong.
"Are you going to hang around to watch the deployment?" Sarah asks Barron.
"Won't be necessary," he replies. "How about if we grab some dinner? You'll be automatically paged if anything goes wrong, along with the other content administrators in Hong Kong, London, and New York!"
It seems contradictory. Web processes break new ground. Experimentation and spontaneity are necessary ingredients in the mixture. Providing the opportunity for new ideas to ignite implies that there is a freedom to experiment, to try new things. But at the same time, an ongoing business needs to preserve fully functioning operations even as hotbeds of innovation within an organization try out new ways to capitalize on web opportunities.
On the surface it sounds paradoxical, but studies of human group performance indicate that people need a combination of freedom and structure to achieve at their highest level. Analyses of high-achieving corporate cultures suggest that balance is key. A culture that is too freewheeling tends to falter because of lack of coordination. It neglects the basic tasks that make an organization work. Because there are no established procedures for most things, some groups excel, while others get into trouble.
At the other extreme, a culture that requires controls and procedures for everything becomes too process-bound, and the result is uninspired performance. Studies suggest that process and structure are beneficial, up to a point. It's best to introduce enough process to ensure that the efforts of individuals and groups coordinate with minimal wasted effort, but refrain from adding process where it hinders more than it helps, or if it prevents the spark of insight from igniting a better way to do things.
This same principle applies to web operations. The ideal is to encourage enough disruptive chaos to allow your organization to capitalize on chance connections, but introduce an infrastructure that builds web assets manageable and reproducible manner.
The requirement for manageability and reproducibility sits alongside the imperatives outlined earlier, which compounds the difficulty. If we follow the other rules but neglect this one, we have an incomplete recipe. If we fail to build an infrastructure that manages the fast-paced, innovative, and diverse environment that we've put together, we'll find that groups trip over one another and productivity suffers. Moreover, in the hyperspeed, talent-scarce world of web development, building a manageable and reproducible practice fosters a happy, productive workforce.