An E-commerce Case Study

The following case study will demonstrate how easy it is for a great web idea to falter. At any point in the life cycle of a web site, the site can fail. Some of the elements that will cause the web site to fail are due to planning and management failure; others are due to economic changes; others are due to technology changes or failure; lack of customer support; bad management; etc. The conclusions in this study are opinion from one perspective and a great topic for discussion.

Scenario
A non-profit Organization had been working with businesses in a natural resources Industry. The Industry had been struggling across the US for many years, and in some parts of the country had basically shut down. The Organization noticed that this Industry was not particularly adept at marketing, especially on the web. So, the Organization decided to create a web site that would function as a place where people could post that they needed to buy or sell raw materials. This idea came as a result of meetings with people from various sectors of the Industry. The Organization was able to obtain a grant to fund the creation and first year of operations of this web site.

This is a classified ad model, similar to buying and selling cars or boats, but within a particular industry. This is not a new model, just a new use of an existing model.

Getting Started
The Organization spent nearly a year researching the idea. A white paper was sent out for opinions. Similar sites were researched, both in the Industry and out. It was found that the sites in the Industry were poorly developed and run, were very localized, or had charges beyond the means of the small businesses within the Industry. There were discussions with technical people about the costs and feasibility. A business plan was written that included the overall functionality of the web site, the market, marketing plan, staffing, and projections. Apparently, it had been planned that the site would be finished by May or June of 2006, but by the beginning of June, the development company hadn't been chosen.

During the planning, a front page design was developed by a graphic designer. An RFP was drawn up that briefly described 25 pages for the site. The type of functionality was described, but not specified in a software plan. Ex. "Events Page List of of highlights of Events, sorted by date. Search fields are Keywords, State and Date." The viewer interface was better defined than either the admin or account interface.

The cost to develop the site was estimated at $25,000.00. The RFP was sent out to about 12 companies to develop the functionality and web site. Only 3 or 4 of these companies responded with a proposal. The other companies indicated that they were too busy or didn't want to take on a project of that size. The companies contacted may have been found by word-of-mouth. The proposals ranged from $20,000.00 to $55,000.00. At this point, the project was almost three months late and development hadn't begun.

In June before the RFP replies had come back, the site owners hired someone with both web development and software engineering training as the "Internet Marketing Specialist (IMS)." The duties of the IMS included:
     - implementation of marketing strategy
     - develop and track web traffic and SEO
     - manage affiliate program
     - manage pay-per-click marketing
     - manage email marketing campaign
     - web site maintenance
     - work with outsourcingfor development and marketing
     - development of customer accounts (sales)

Choosing a Software Package
On analyzing the business plan and RFP, the IMS discovered that the functionality was just a web classified ad system. This employee suggested also considering an off-the-shelf software product to shorten the development time. The IMS developed a spreadsheet of basic functions required, although many of the details were unclear from the existing documentation, and there was no software plan. Then about 12 OTS products were compared to the list of functions.

After the list was narrowed down to 4 or 5, references were followed and calls to the software companies were made. Two of the companies had online references to sites that no longer existed or gave very unflattering comments about the products. One of the references even made a suggestion of a different piece of software. The highest bidder gave a presentation of how the software might look. While the presentation was very professional, many of the features tauted as "our software" were actually modules available in the CSharp programming environment. However, this was only apparent because the IMS had programming experience. All the references (4 were called) for AlphaPlex and their MarketGrabber software were positive.

Another consideration was the programming language. One short-listed product was written in Cold Fusion, which would have been technically difficult to manage at all in-house with the skills available. A manager of one of the companies that submitted a proposal said that "no one uses Cold Fusion any more." This is not true, although it is more commonly used by very large web sites. The IMS was proficient in PHP and MySQL, but only had minimal experience with either ASP or .NET. According to the business plan, all programming was to be outsourced, but already having someone in-house with programming skills was a consideration.

After two weeks of research, the MarketGrabber product was selected. The business plan called for two level of ads, events and news. MarketGrabber had several packages, none of which satisfied all the requirements. However, the Classified and Community packages together contained all the basic requirements. The problem was that the packages were meant to run separately. Although there were quite a few functions in common, each module had its own complete set of files. Therefore, there was a lot of code duplication. So, while the site was supposed to show both levels of ads on a page for each of the product categories, the software was designed to show the two levels of ads on different pages. Quite a few customizations to the packages were required to join both Classified and Community ads be available in the same lists by topic or other query.

An example of a problem with the customization:
What if Product Category A had 50 ads in Classified and 10 ads in Community. If a page is supposed to show up to 20 of each type of ad and then show a link to more pages, should the same 10 Community ads show again on pages 2 & 3? This was especially problematic if the Community ads were on top, because it would look like the same page over and over. Since clients were to pay more for the Community ads, they expected top placement; so, the ads couldn't be mixed together.

The price of the original packages was something less than $5000.00. With the customizations, the price for the initial set up came in at less than $10,000 - much less than the budget and much less than the top bid of $50,000.00

Issues

1. No Software Plan
   The RFP only outlined the barest of page functions. The off-the-shelf product selected had great basic functionality, but the user interface was very bare. Two years later, management stated (correctly) that the user interface was "stupid," but there was no definition at any time of what it was supposed to be. This should have been dealt with during testing.
2. Unrealistic time line
3. Duplication of code in software packages.
4. No Testing Time
5. Terminology issues - "classified ads"
6. Lack of experience with the complexities of the "backend" of such a system.
7. "Implementing a marketing strategy" was really a sales position. In fact, the IMS was just a sales position, the other jobs were supposed to be done in non-existent time.
8. Unrealistic staffing plan for this type of site.
9. Unrealistic expectations that once the site was functional, it was finished.
10. No allowance for readership needed for advertising sales - no targets, workplan or budget for circulation development.
11. Market buy-in lag
12. Time was spent working on the affiliate program at the beginning that should have been spent on marketing.
13. Not being honest with the board about the realities of the costs of developing and ramping up such a site. These realities were part of the learning curve, but the new information was not given to the board.
14. No implementation, plan or budget for site maintenance.
15. There was no review of the documentation to match the product or procedures to the documentation. The documentation was not reviewed periodically to realign either the documentation or the product and procedures.