6.13 Improving the Business

Whatever the planning, or the experience it was built on, the sad fact is that few sites perform quite as expected. Nonetheless, while before you had just theory and other sites to go on, now you have your own site with your own visitors to turn into ever-increasing customers. Not only is traffic generally low in the opening months, but so, very often, are conversion rates. These are the things (in no particular order) you'll probably find yourself doing:

1. Improving landing pages. You'll get an idea of customer's interests as the popularity of certain pages emerges. Make sure you pick up that interest, adjusting the site structure if necessary.
2. Building non-reciprocal links.
3. Checking your email capture system is working properly. No point in inviting newsletter subscriptions if you can't a. collect all email addresses and b. learn something from their responses.
4. Reviewing site security.
5. Testing sales and other copy. You don't have to wait until traffic builds to assess results with split-testing software.
6. Checking the site navigation. Is it absolutely impossible for the visitor to get lost, however dim-witted?
7. Double-checking the ordering process with third-party guinea-pigs. 30% of sales are commonly lost here.
8. Finding affiliates to market your products.
9. Understanding your customers: who they are and what they are really looking for.
10. Join a business affiliation scheme to give credibility to your site.
11. Improving site design to attract preferred customers and improve conversion rates.
12. Creating trust through newsletters or ebooks.
13. Extending advertising on pay-per-click search engines, comparison search engines and eBay sales.

Landing Pages

Landing pages are where potential customers enter your site, and these are the points to check:

1. Pages immediately give visitors what they're looking for: catalog, special offer or information.
2. Identify what you expect customers to do, and design the page around that, removing obstacles and distractions.
3. Highlight benefits with graphics and case-histories.
4. Emphasize the special offer.
5. Increase confidence with guarantees and transaction security features.
6. Provide location and contact information: a real business with premises and people.
7. Feature phone support or live chat.
8. State the returns policy clearly.
9. Add testimonials.
10. Use customer tracking to monitor behavior.
11. Experiment, checking with split-testing.
12. Streamline the check-out route, the shorter the better.

Expanding the Business

Suppose all is well, and results are comfortably within expectations. You'll be wanting to build on your investment, expanding the website as sales suggest. The usual advice is not to be over-optimistic in your original plans, or pay for what you may not need. Nonetheless, the site should have some provision for the development envisaged by your business plan. If it's simply a matter of increasing the product line, then you can add more website pages or upgrade the storefront program. But if you're moving into another dimension with database storage of product information, or branching into new product lines altogether, then a proper plan is needed. Requirements vary, but the general advice in these circumstances is:

1. Develop a new site separately, under a different domain name if necessary, so as not to interrupt the smooth operation of your present site. Test extensively before directing visitors to the new site, or replacing the old site with the new.

2. Consider employing a professional web design company. You'll have a clearer idea of what's wanted by now, and much of the time-consuming and expensive consultation can be side-stepped — in logo design, overall look and functionality. Aim for continuity: some aspects should stay in the new design to extend the trust you've built with your present customers: i.e. go for a makeover but not a total change.

3. Employ a full-time webmaster if you don't already have one.


Webmasters are responsible for the overall look and content of the website, close to the editor's role in a newspaper or magazine. The analogy implies responsibility and experience, and good webmasters can indeed turn their hand to graphics, copy writing and programming. More importantly, webmasters are the creatures who draw all aspects of the business together, and turn your ideas into web reality.

Good webmasters are therefore unusual creatures, and perhaps only the best really shape up. Any IT Recruitment Company can find you candidates with the usual skills in HTML, Javascript, VBScript, in Photoshop and in site build programs like Homesite or Dreamweaver. The stronger candidates will also have some programming or scripting experience — Perl, ColdFusion, ASP, and VB — and be familiar with Oracle, SQL Server and/or MySQL databases. But your needs will go further than technical expertise. You'll want the outstanding individual who can play a key role in liaising between different departments and objectives.

It's a two way street. The successful candidate should be able to look at your site and suggest immediate improvements, estimating time and cost involved. And you should be able to spell out exactly what your future plans entail, with some idea of budget available. Supply exceeds demand, but webmasters are still ambitious and creative people who want to be associated with the best going.

In summary, a good webmaster should be able to:

1. Iron out any wrinkles in the site left by the designers.
2. Work with you in continuously improving content and layout.
3. Promote the site through the search engines or work with an outside optimization company.
4. Monitor viewer behavior and sales, translating the ideas of the Sales and Marketing departments into effective web design.
5. Supervise any IT contractors upgrading site appearance or functionality.
6. Continually check and improve site security.
7. Keep you apprised of Internet developments important to the business.
8. Translate the traffic statistics into meaningful measures for management.

Maintenance by the Web Design Company

Nonetheless, you may decide that budget and current needs do not justify a full-time webmaster. Most web design companies will offer to maintain the site for you, offering very reasonable terms. You should accept, but get details clearly spelled out in the first contract. Web design companies make their money in building sites, not in maintaining them. They are often reluctant than do more than add the odd page and keep the site running. And this is understandable. The $100/month or whatever maintenance fee only buys a few hours of work, and a major change — adding a database or entirely reorganizing the page layout — is clearly a new contract.

Understanding the Customer

You've planned and built the site about some marketing model, and visitors are buying. It's a success. Why bother with customer research when you're more than busy fulfilling orders?

Because it's the first law of marketing. Understand your customers — what they are looking for, why and with what degree of success. Recent surveys have found that emerchants can now generally quote their conversion rates, but many still don't know their shopping cart abandonment rates. Still less do they know these essential matters:

1. Conversion rates applying to: customers from natural search engines as opposed to pay-per-click search engines.
2. Customers from individual search engines, which vary considerably.
3. Customers grouped by occupation, age, country, or disposable income.
4. Customers drawn by affiliates, banner ads, press releases, site awards.
5. Visitor behavior at each stage of selling process, i.e. page to page through the site.
6. Profits and ROI on advertising spend grouped by: particular keywords and ad copy on pay-per-click search engines and sponsored listings.
7. Influence of site design, page copy and ease of navigation on sales.

You're flying blind without this vital information, and even small changes can have dramatic effects. A study by Marketing Experiments analyzed two marketing services and found that, while one was very profitable, a second with over 2,500 customers was in dire trouble, losing 40% of its subscribers monthly. Yet the reasons were very simple. The second company had neglected to:

1. Identify their proper customer base.
2. Use descriptions and metatags to filter out unwanted customers.
3. Screen the market with prices used as both an attractor and barrier.
4. Make their copy sufficiently specific.
5. Design their home/landing page to appeal to their desired customer.

Examine the Site Traffic Statistics

Your first task is to look at the traffic statistic, preferably weekly, but at least monthly, to answer these questions:

1. Time spent on site (wrong visitors if too short: adjust copy and keywords).
2. Ratio of pages viewed to visitors (visitors confused or not interested if low).
3. Ratio of visitors to unique visitors (encourage repeat visitors: they buy).
4. Visitor country of origin (improve marketing spend in some of these?)
5. Search words and phrases used to find the site (coincide with the planned keywords?)
6. Importance of (i.e. referrals by): search engines and directories (tweak the pages to improve ranking as necessary).
7. Affiliates (are they really working for you: can you help them?)
8. Pay-to-click search engines (acceptable conversion rates?)
9. Reciprocal links (which sites are helping you?)
10. Points of entry (the pages that greet your visitor: are they doing their job?)
11. Points of exit (pages where visitors give up: why?)
12. Typical visitor routes through the site (is this the pattern you want?)
13. Percentages proceeding to sales pages (improve site navigation if necessary).
14. Percentage actually purchasing (calculate the various conversion rates: acceptable?)

Customer Tracking

The next stage involves customer tracking software or services to model visitor behavior, most particularly with funnel analysis, applied to traffic from:

1. Natural search engines.
2. Pay-per-click search engines.
3. Individual search engine referrals (days, hours and countries).
4. Affiliates, banner ads, press releases, etc.


You may wish to run a survey: many companies will set up one for you and provide helpful advice.

Blogs and Newsletters

Ideally, you want to make friends of your customers since it's usually by repeat orders that a business thrives. Hence the importance of blogs and newsletters, which create a community where the customer's voice can be heard and the company get its message across.

Bulletin Boards

A further step towards creating a community are bulletin boards, interactive chat boards or even a classified ads service for subscribers/ customers. All require moderating, and take valuable time to supervise, but are worth considering in these circumstances:

1. You sell high-value software.
2. A user's forum can create a sense of confidence in your products and drastically reduce the time technical support spends on answering 'dumb questions'.
3. Your service is community-based anyway: hobbyist sites, trade associations, specialist bookstores.

Customer Emails

Some customers are never satisfied, but the complaint email or telephone call can be turned to your advantage. In their unthinking ways, visitors may really have stumbled over some design shortcoming. It's often possible to convert an irate purchaser into a faithful customer, moreover: 'Dear Mr Stevens, We're truly sorry you've had this experience, and of course are making an immediate and full refund. In this connection, you may like to know. . .' Keeping emails, and going through them at convenient intervals, may indeed be your best way of sensing what your customers are really looking for. Some will suggest new opportunities, or at least give you that treasured testimonial. Complaints are less a nuisance than a selling opportunity.


Most companies manage their affiliates through third party providers like Clickbank and Commission Junction, but studies by Market Experiments.Com suggest that homegrown partnerships can deliver better results. Very large numbers of companies commonly sign up as affiliates — thousands or tens of thousands — but only half of these make a sale, and only the tiniest fraction contribute significantly to your bottom line.

Improving Site Design

Marketing Experiments found that problems causing a serious drop in revenue included:

1. Site was not focused around one clear objective.
2. No clear customer problem and its solution were communicated.
3. Homepage did not 'hook' attention in the first five seconds.
4. Sufficient incentive was not given for every action requested.
5. Prospect was not continually involved in the purchase process.
6. Credibility indicators were not subtly placed throughout the site.
7. The sales copy did not convey integrity and accuracy.

Sales Copy

How long should sales copy be? The results of MEC's tests were:

1. Both long and short copy have their strengths, but expensive products do better with long copy.
2. Sales copy and product specifics are separate matters. For decent conversion rates, customers need to be given appropriate sales copy, sent through a page providing specific information and then transferred to the shopping cart. Shortening the process by removing the specifics page drastically reduces sales.
3. Good copy outperforms poor copy every time.

Awards and Credibility

Dozens of organizations provide awards, but are they worth anything? Is the time spent polishing and submitting the site properly rewarded in better sales? MEC's findings were:

1. Over 600 agencies give awards, several of which can be won if the site is properly designed.
2. The agencies themselves send negligible traffic, but
3. Awards combined with strategic press releases do enhance confidence and therefore sales.

Order Form

The most telling fault of ecommerce is shopping cart abandonment at present, running at 25% to 32% and costing $6.5 billion/year in lost sales. Research by BizRate and The NPD Group found that, of the 'almost customers', 39% did not purchase the item at all, 26% purchased the product from a competitor, 17% made their purchase off-line, and 18% returned to the site to make the purchase at a later date.

Much more under the control of the emerchant was design of the shopping cart/order form. Even a simple shortening of the order form increased sales by 18%, and MEC went on to list 29 ways of improving sales in the 18-45% range.

Ezine/Ebook Experiment

In their ezine/ebook experiment, MEC built an ecommerce store ($7,500), created an associated free ezine ($3,500), and reformatted the ezine material as an ebook (additional cost of $1,500). They wanted to know how the offer of an ezine and ebook would affect traffic and sales at the commerce site. The results, very briefly, were these:

1. The ezine increased traffic considerably, from 12,000 page views/month to 70,000 page views/month over a 6 month period. It also increased retail sales six-fold over the same period.
2. In fact, two ezines were produced, one appearing weekly and the second fortnightly. The second ezine was more successful in increasing page views at the ecommerce site (276% as opposed to 165%, both over on the day following launch of the ezine).
3. The second ezine was used in an attempt to convert free subscribers to paid subscribers. 17% of viewers clicked through the offer page and 3% of these went on to convert to paid subscription. This 0.5% overall conversation rate was for one ezine issue only, but was worth $289 annually. If the ezine could win 50,000 free subscribers, and could entice 60% of readers to find the offer page, then the annual value of paid subscriptions would amount to $268,000.
4. The second ezine experiment was repeated, this time offering an ebook in place of the ezine. Conversion rates were then 8% rather than the original 3%, a 267% increase over the ezine and worth $771 annually. Over a ten year period, the annual subscription to the ezine was worth 4 times more than the one-off ebook subscription, of course, but MEC recommended offering both as a relatively cheap and effective way of increasing sales.

MEC were not wholly clear whether the ezine was part of the site, i.e. a true ezine, or was mailed out, when newsletter might be better term. The math was also somewhat simplified, and many assumptions are made. Nonetheless, the potential of ezines/newsletters was clear. A newsletter could be made to pay for itself if only modest conversion rates to paid subscription were achieved. The way was then open to reward subscribers with sweepstakes, cash prizes for 'letter of the month', listings of reader's 'offers and wants', etc.


MEC submitted to 36 directories, newsgroups and review sites and found, over 30 days, that:

1. Many directories were a waste of time: over-hyped and poorly maintained.
2. Newsgroups generally had membership too small to be worth bothering with.
3. Review sites gave good returns when the submitted ezine was actually featured.
4. 120 new subscribers were gained for an expenditure of $600.
5. A few directories were well worth the effort, and should be used.

Using Advertising Spend More Effectively

A common measure for advertising campaigns is the ROI, the return on investment: how many dollars you get in profit for each dollar invested. The last is not merely ad spend, of course: there are salaries and overheads involved in research, planning the campaign, monitoring and writing up results. But to make this simple, consider a real test case reported by MEC. They spent over $24,000 on click fees in three campaigns, and found that by far the best result (in ROI and marketing cost) was achieved with the low bid price of $0.08. Nonetheless, the best profit was achieved with a click-through costing $0.50. ROI and profit are different entities.

Using Cheaper Keywords

Some keywords on Google and Microsoft are becoming prohibitively expensive. MEC looked at the case of an Internet supplier of telephone parts whose preferred keywords cost $1.49 to $1.52, when conversion ratios allowed no more than an average $0.25. MEC's solution was to use 170 keywords that were low-cost, less precise but nonetheless effective.

Using the Smaller PPCs

MEC found that the smaller ppc search engines can bring excellent returns on small investments, even though the conversion rates are much lower and — not taken into consideration here, but obviously important — the profits may be small for the management effort involved.

Selling on eBay

Sales via eBay are more difficult to model because eBay charges an insertion fee whether the item sells or not ($0.30 to $4.80), plus a commission on the final selling price (5.25% of the first $25 plus 2.75% of the remainder, up to $1000). In general, eBay is an attractive alternative to the ppc and price comparison search engines only if you can:

1. Price competitively, and maintain a good profit margin.
2. Offer superlative service: customer feedback is shown.
3. Devote considerable time to this aspect, or hire an expert to do so.

Price Comparison Search Engines

Internet shoppers are canny creatures, and increasingly make use of the price comparison engines: DealTime, AOL, My Simon, NexTag, BizRate, PriceScan, PriceWatch, Price.Com, CNet and PriceGrabber. Should you advertise with them?
In comparison with Google AdWords and (the then) Overture, MEC found that price comparison engines:

1. Require more competitive pricing
2. Take longer to register with
3. Can send larger volumes of traffic
4. Have lower costs per click
5. Higher conversion rates.

From varied material provided by MEC, these rough comparisons can be made:


Oct 2002



2002 Traffic

Conversion rate

Yahoo Shopping




MSN Shopping




AOL Shopping




































AOL Shopping




My Simon












Price comparison engines are clearly worth sustained attention.


What's the best price? As always, the problem is to find a price high enough to get good margins and low enough to achieve wide sales.

MEC demonstrated a practical approach. The product in question was an ebook on coping with depression, and MEC's first step was to identify the competition.

1. They found three books being poorly marketed through mainstream online booksellers, i.e. 'lost in a sea of digital generalization'. They therefore focused on child psychology, and found a site offering books, articles and videos. The site was being marketed through the ppc search engines, paying $0.17 per click for "bipolar" and $0.05 per click for "bipolar children".
2. Encouraged by that finding, their next step was to locate other such sites and study the keywords being used in the metatags. Depression, mental health, psychology, child depression, and therapy were the most popular, but these keywords had high bid prices on the leading ppc search engines, even for a third place ranking.
3. That being the case, MEC identified keywords with more acceptable costs per click: bipolar disorder, bi polar, child psychology, bipolar and schizophrenia.
4. MEC then checked to see that a significant demand existed for the ebook in question. Sending traffic to a test page with the cheap keywords, MEC achieved 648 click-throughs for an average cost-per-click of $0.1. Of those 648 click-throughs, 84 clicked a dummy order link. Some would no doubt have abandoned the purchase process, but the initial 13.0% conversion ratio was sufficiently positive to proceed to pricing experiments.
5. Five keywords were used in the same ppc search engine over three days. The $9.50 ebook attracted fewer orders than the $14.95 ebook: customers obviously felt that something as cheap as $9.50 wasn't worth the money. Secondly, though the $24.95 price attracted fewest orders, it gave the highest revenues.

Pricing therefore begins with an educated guess, is followed by testing, and then moves to the highest price the marketer can get away with.

Shipping Charges

MEC also looked briefly at shipping charges. Though customers are attracted first by price, they also take into account the shipping charge. Indeed, studies by Jupiter Research showed that 51% of customers shopping off-line did so to escape the shipping charges. Should you offer low cost or free shipping, therefore? And, if so, how do you make it pay for itself? MEC's conclusions were:

1. High shipping price is a tactic that works on eBay, particularly for one-off items.
2. You can also get away with higher shipping charges on estores if you are not aiming for repeat business.
3. In all other cases you'd be well advised to:
      a. keep your combined charges competitive, perhaps adding a 'priority service', or
      b. offer free shipping for orders a reasonable amount over the average — one which seems good value, differentiates you from competitors, and where the 'free' is offset by increased sales, perhaps from discounted items. You can find this point by modeling possible scenarios in the light of your own experience, and
      c. monitor pricing experiments carefully.


1. What often has to be improved in the first few weeks of an ecommerce site going live?
2. How can you better understand your customers?
3. Why would you employ a webmaster, and how interview?
4. List the points to check on landing pages.
5. Describe MEC's ezine/ebook study. How could the results be useful?
6. Why is it worth experimenting with price comparison search engines?
7. Suggest some ways of using the marketing spend more effectively.
8. How would you find the optimal price for an ebook?
9. When should you make money on shipping charges?

Sources and Further Reading

1. Marketing Experiments. Extensive studies in page optimization.
2. Business statistics. Bizstats. Basic but free.
3. Practical Ecommerce. Much useful advice from those actually running ecommerce sites.
4. Small business benchmarks. Australia Taxation Office. Used to detect tax avoidance.