Managing Your eCommerce Environment:
What you need to know to compete successfully on the Internet
Summary
eBusiness and eCommerce were buzzwords of the late 1990s, and they show no sign of diminished publicity in 2001 or for the forseeable future. Using the Internet as a trading platform for goods, commodities and services has become one of the biggest growth areas in the world, with IPOs and paper millionaires now daily news.
But there is more to eCommerce than simply having an idea, setting up a website, and watching the share price rocket. More Internet ventures fail than succeed, and those that succeed do so because thought has been put into the strategy. Just as a shop that is closed or constantly out of stock when you want to buy something will very rapidly lose your business, so an eCommerce site that is frequently down, or that mis-processes your order will not succeed in luring you back for a repeat visit. Nor will a site that crashes when too many people access it. And those repeat visits are the key to success.
Ensuring that your website is run as a mission critical system, with such factors as load-balancing, failover, server-mirroring and auto-repair all in place, is vital; overlook these things and your venture will almost certainly fail. When you consider that some sources predict that eCommerce will top nearly a trillion dollars by 2002-and believe it or not, that is a conservative estimate-there is clearly much to be gained by securing your company's own eCommerce future. But, unless you take care to ensure the robustness of eCommerce management, that just isn't going to happen.
This paper looks at the development and future of eCommerce, and the steps you will need to take to make sure that your systems and systems management are robust enough to face-and win-the challenges of eCommerce in the 21st century.
What is eCommerce?
It is virtually impossible to open a newspaper or a magazine now without seeing the latest Internet success story plastered everywhere. New paper millionaires are being announced daily, and the entire Internet phenomenon has spawned its own vocabulary: clicks and mortar, dot com companies, eCommerce and eCommerce, to name just a few. But for every success story that makes the news, there are possibly as many as 1000 Internet ventures that collapse without a trace. Remember, 95 per cent of all new businesses fail within the first year, and it is far easier to start an Internet company than a traditional bricks and mortar company. Don't be misled: casualties are high and the Internet takes no prisoners.
So, what distinguishes a success from a failure in this new trading arena? First, we have to look at what is meant by eCommerce. There are actually three distinct types of eCommerce, with three subtly different business models. (Note that the terms eCommerce and eCommerce are now largely synonymous, although there was a time when they actually were different: eCommerce concerned itself with reaching out to consumers, while eCommerce meant business-to-business trade. This paper will use the terms interchangeably.)
Business-to-business eCommerce is actually the oldest form of Internet trading, with its roots in the more traditional mainframe to mainframe transaction processing of the '60s, '70s and '80s. The majority of Internet transactions are currently business-to-business. However, it is business-to-consumer eCommerce, also known as e-tail, that is hitting the headlines, with companies such as Amazon.com and lastminute.com processing phenomenal numbers of orders, threatening almost unbelievable IPOs and generating considerable press coverage. The third type is consumer-to-consumer eCommerce, where the Internet is used as a platform for consumers to trade among themselves, with the website offering support and services. Perhaps the best known example of this is eBay, the online auction site, which got a welcome publicity boost last year when human organs and eggs were put up for auction. Although it is unlikely that consumer-to-consumer eCommerce will ever generate the same revenues as the other two types, it is still an important player in the eCommerce market, because it has potentially more power to pull in consumers than the others.
Within these three models, there are a number of generic types of trading. The simplest is to publish your catalogue or inventory online. This breaks no new ground. But when you go one stage further, you open up a whole new world of commerce: allowing people not just to browse your stock online, but to buy online as well. The first widescale implementations of eCommerce were exactly this: the website simply became an electronic version of the store. Purchase requests were emailed in, money was sent by snail-mail, and the orders were fulfilled. At this point, the fledgling eCommercees hit their first hurdle: how to integrate their existing order fulfilment architecture-frequently a back-end mainframe system-into the new world of the Internet. As we will see, a number of IT companies specifically address this problem. Once the integration hurdle was cleared, many traditional marketing tools were brought to bear-advertising and PR are as applicable to eCommercees as they are to their old-fashioned cousins.
However, possibly the biggest difference between bricks and mortar and dot com has been the revenue stream. Even Amazon.com, cited as the greatest Internet success story of all time, has yet to make a profit. Its value is intrinsic: Amazon.com is worth billions because the people who count-financial analysts, stock marketeers and so on-believe that it is. The traditional profit and loss model breaks down where a lot of dot coms are concerned. But where does the money come from? All types of eCommerce fall into one of two camps. 'Clicks and mortar' companies are the traditional companies that have launched an Internet arm. The success of Egg (part of Prudential), Smile (part of the Co-op Bank), barnesandnoble.com and JCPenney.com is in part due to the backing of the parent companies, which are prepared to bankroll them until they become profitable. The true dot coms are those without traditional backing. For them, money has to come from somewhere and the most common source is venture capital. But usually neither type of enterprise depends on the profit and loss model to keep going. There is also the question of overhead: a traditional store needs a warehouse, store space, cashiers, floor personnel … A dot com needs none of these. Orders can be processed without the any physical inventory being on hand; once the order comes in, the dot com simply directs it to a third party for fulfilment. The rise of the dot com has introduced a number of new business processes-as well as a number of serious problems, as we shall see later.
But where did eCommerce come from?
The development of eCommerce
Surprisingly, eCommerce began nearly 40 years ago. In the 1960s, a set of standards was drawn up enabling companies to format their data in such a way that it could be read by other companies. This information could then be transmitted across the available communications infrastructure, such as telephone or computer-to-computer across an early network. Once the infrastructure of the Internet was in place, it soon became obvious a whole new means of commerce was available, and eCommerce was born. Internet business and eCommerce follow a set of rules that differs significantly from the rules governing traditional commerce. Whereas traditional businesses adhere to standards and practices established hundreds of years ago, the guiding principles of much of the eCommerce world are of comparatively recent invention.
The future of eCommerce
The number of eCommerce companies is growing exponentially, with no end in sight. However, this is not to say that everything will remain the same. Last year, everyone was talking about floating dot com companies; in 2000 this will begin to happen more and more frequently. Doomsayers often liken dot coms to the Emperor's New Clothes: they retain their market valuation only as long as people have faith in their prospects.
What is almost certain to happen is the survival of the fittest, just as in traditional business: those companies unable to satisfy their customers will go to the wall. Dot coms will cease to be valued by clever guesstimates and more by the strength of their revenue stream. Many analysts believe that, in no more than ten years, the dot coms will be as respectable as their bricks and mortar forbears, adhering to the same business models and processes, dealing with the same obstacles. But in today's market there are numerous obstacles peculiar to the Internet-obstacles that can potentially spell doom for an unlucky dot com.
Obstacles to eCommerce:
Integrating old and new technologies
As we have seen, many of the so-called 'clicks and mortar' companies-Internet arms of existing traditional companies-had early, and often disastrous, encounters with back-end integration. Many of these companies still store the majority of their corporate data on legacy systems, such as mainframes, rather than on modern PC-based networks. There are many reasons for this: investment in the data structures, cost of replacement, or simply familiarity. But for the eCommerce to succeed, the back end must be able to talk to the front end. This was understood by a number of companies early on. IT companies that provided basic screenscraping-capturing mainframe information and putting it into a PC or web browser format and back again-saw that this could be used to firmly marry the two sides. Why is this so important? Web transactions happen in real time. Therefore order entry, fulfilment and inventory checking must happen in real time as well. There is no point in buying a car over the Internet, paying for it, and then discovering that someone else bought the last one five minutes previously, but the mainframe hadn't been updated. Result: an unsatisfied customer. Companies such as Attachmate have made major strides in ensuring that the old and new technologies work in co-ordination, enabling real time updating of the back end, so there is less danger of losing transactions in the gap between old and new.
Security issues
In the first quarter of 2000, denial of service attacks on prominent websites, and theft of credit card information from others, focused world attention on the necessity for more effective monitoring of the activity and security of all computers connected to the Internet. Indeed, IT security in general, and Internet security specifically, is one of the fasting growing sectors of the IT industry. The technology now exists to prevent all but the most determined of hackers from accessing confidential data: encryption, firewalls, anti-virus software and even software built into the website itself can ensure that an e-transaction is as secure as any other type of transaction. Companies pursuing eCommerce often focus on creating a pleasurable and efficient customer experience-efforts vital to attracting customers. To keep the trust of these customers, eCommercees must make an equivalent commitment to implementing available security technologies, and pushing for the development of further safeguards.
No one is stating that the Internet is hack-proof, but then again, no house is burglar-proof. Security is actually the management of risk. The onus is on eCommerce to ensure that customers' trust is never compromised. A breach of trust-like a loss of reputation-is a very difficult thing to fix.
Legal issues
The legal position of Internet business is actually a two-edged sword. On the one hand, the worldwide nature of the Internet means that tax considerations become more fluid. Currently, some eCommercees enjoy the advantage of operating in jurisdictions with low or non-existent taxes. On the other hand, eCommercees must also take care that they do not incur fines or legal expenses through the sale of items forbidden in the jurisdictions of their customers. And this is nothing compared with the logistic nightmare that is copyright law across different countries-not a topic we can cover in a paper of this scope.
Cost
One factor inhibiting eCommerce may soon be irrelevant-cost. Currently, it costs money to log onto the Internet, and in some countries this cost is considerable. For business-to-business e-companies, this isn't too much of a problem. Leased lines and high-powered ISDN connections are common for them, and the extra traffic generated through eCommerce isn't problematic. It is the business-to-customer sites that have been hit by this. Various surveys and reports suggest that eCommerce will double or even quadruple if Internet access becomes free, and in such countries as the UK, recent moves by British Telecom and other telcos suggest that unmetered, low-cost high-speed access is just around the corner. This is one obstacle that might soon be a thing of the past.
Bad Design
A whole industry of social engineers, ergonomicists and psychologists has sprung up around the age-old question of 'what makes people want to buy?' The layout of supermarkets has been refined to an exact science, with the location of certain items dictated by the human mindset. No one would consider opening a shop and just throwing the stock around at random. Navigating the web is just as much a science as store design, but far too many sites appear to have been constructed without this in mind. For eCommerce really to take off, websites must be accessible to everyone, not just the technically proficient. One of the largest growth areas for the PC market is currently the 'silver surfers': financially secure people over 40. They have the money to spend over the Internet, but many of them do not have the technical expertise to find their way through a complex hierarchical website. Their money will be one of the vital factors in developing eCommerce.
Downtime
Clearly, there are obstacles to eCommerce, just as there are obstacles to any business venture. However, the Internet has brought with it some problems that are specific to online trading. Perhaps the most important is downtime. In a traditional bricks and mortar company, the consequences of being prevented from trading are fairly obvious: loss of earnings, loss of reputation, and so on. Although these factors are equally applicable to eCommercees, they are considerably more severe. Traditional shopping is a fairly passive exercise. If one particular store is closed, customers may go to a rival store, but their first choice will still be there when they return another day, enticing them to try again. Customers have to be far more active in an eCommerce environment: they have to type or select the URL, navigate through the website, and then make their purchases. Even one instance of downtime can be enough to erase the memory of the URL from the customer-if it didn't work then, why should it work now? There has been considerable research into what makes a site 'sticky'-that is, what makes people come back again and again. And one definite conclusion was that frequent downtime is one of the best ways to lose stickiness.
There are financial implications as well. Dell Computers broke all the rules-and changed the face of the reseller market forever-by offering its PCs directly to customers online. Its website currently sells $5 million of equipment per day-imagine the consequences of even an hour's downtime. If Amazon.com was felled for an hour, it would lose an estimated $70,000. We're talking real money here. And it's not just the front end. As we have seen, integrating the customer-oriented HTML pages of the web-site with the legacy system back end is vital for trading in real time. If the back end fails, there is the potential for mismatches between online transactions, as well as lost orders and severe data corruption.
Together, these possibly form the most crippling problem any eCommerce company can face-loss of reputation. Reputations take a long time to forge, even in the speeded-up world of Internet time. But they can be destroyed, beyond all hope of repair, with just a few hours of downtime, a couple of lost orders, or a labyrinthine website. And once faith is lost, it isn't long before the venture capitalists and investors take notice.
So…
For eCommerce to be as successful as the predictions, all of these obstacles must be addressed. Some of them can be solved more easily than others-cost can be addressed by a free Internet access, design by hiring good web designers-but others are more problematic. The most debilitating obstacle is loss of reputation. And if that is caused by downtime, many eCommercees have only themselves to blame.
Maintaining your reputation through e-management
E-customers expect the same level of service from their supplier as traditional customers, if not better. They expect the website to be open when they expect it to be open-and over the Internet, this isn't nine-to-five. They expect the product they want to be available when they want it. They expect transactions to be processed cleanly, quickly and securely. They expect the goods to be delivered when the supplier says that they will be delivered.
The moment the website fails any of these expectations, customer trust starts to erode-very quickly.
But it is horrifying to discover that many eCommercees put the minimum of thought into these factors.
Many believe that a catchy URL will make them millionaires, and launch without any serious consideration of the risks involved. To take a fictitious example, if noplanning.com launches with 150,000 registered users, it would be wise to ensure that the web server can handle 150,000 hits all at once. If the site has a maximum capacity of 100,000, and it fails when everyone accesses it simultaneously, then someone hasn't done their homework. This calamity can be so easily avoided. Technology exists to balance web traffic loads across multiple, clustered web servers. Companies such as HydraWeb offer the Hydra, Microsoft offers Network Load Balancing. Those 150,000 hits can be channelled across different web servers, ensuring that everyone can have simultaneous access. But did anyone in noplanning.com consider stress-testing the web server? Of course not.
Assume that noplanning.com has solved its load problem. There is then the question of refreshing the website. They could possibly switch to a mirrored server when use of the primary web server is interrupted, but the chances are they haven't got one. They bring the site down at three a.m. when no one is watching … apart from the millions of people in different time zones. This downtime could have major implications if noplanning.com wants to reach a multinational audience.
Time for noplanning.com to buy another server and mirror it, using an application such as Legato's StandbyServer (formerly Vinca's StandbyServer) or Microsoft's Cluster Server. With a mirrored server, failures or downtime in the primary web server can be immediately covered by switching to the mirror.
But noplanning.com isn't out of the woods just yet. The web servers are at the heart of a wider network, which entails a whole host of potential problems: network crashes, server failures, power failures, data corruption … If any of these disasters happens while noplanning.com's website administrators are available, there is a good chance that the damage can be repaired without too much fuss. But at three a.m., when the automated refresh is happening? By nine a.m., when the IT staff members arrive, noplanning.com will have been down for six hours. Six working hours somewhere in the world. Once again, lack of foresight could just have cost noplanning.com its reputation, and, once again, it could have been avoided. Products such as Heroix RoboMon and RoboER offer emergency diagnostic and repair facilities, even when no one is there. With applications such as these in place, noplanning.com's network could have been monitored effectively, preventing many problems before they happened, and repaired automatically in the event of trouble-with minimal downtime.
However, noplanning.com has noticed that all of these products come from different vendors. Surely there must be a suite available? Actually, there may be. But buying it might be too foresighted for noplanning.com. According to analysts the Gartner Group, 'the framework as a technical solution has failed.' It took many years for the concept of end-to-end hardware solutions from a single vendor to be blown away by a 'best of breed' approach. Unfortunately, the framework concept is still pervasive as far as software is concerned. Many companies are beginning to believe that the best way to protect their networks is not to go for an all-encompassing single solution in a box, but to look for an individual solution for each particular problem. Systems integration is no longer the nightmare of protection clashes and incompatibilities it once was-more standards are in place to ensure compatibility across and within platforms.
For an eCommerce, the website is mission critical. Lose it and you potentially lose your reputation-and your business. When you are furnishing a house, you don't usually buy all the furniture, electrical goods and kitchenware from the same manufacturer, so why should a network rely on protection from a single source? To ensure that your site is properly protected, go for best of breed. Best of breed server mirroring, best of breed load-balancing, and best of breed diagnostic and repair software.
Conclusion
There is no doubt that dot com companies are going to be big news for the next couple of years. But for every success story such as eBay, lastminute.com and Amazon.com, there will be hundreds, if not thousands, that suffer the ignominy of failure. A large proportion of these sites need not fail. With detailed preparation and planning, and acknowledging the obstacles to be faced, your business could very well be the new Amazon.com. An eCommerce is still prone to some of the same problems facing traditional companies, with loss of faith the most crucial. For any self-respecting website, maintaining its reputation is key. And ensuring that it's robust and available 24x7 is key. But don't look to a suite of products to solve all of your problems. For each part of your eCommerce, buy best of breed for the maximum amount of protection and peace of mind.
Written by Craig Hinton
