Primary metric: conversion and abandonment

Secondary metric: offline activity

If you’re running a transactional site, you need to constantly compare different site designs, offers, and pricing strategies to find the ones that have the highest conversion rates. Page views mean nothing unless you can turn visitors into buyers.

If the end of your transaction happens offline—talking to a sales rep, getting contact information from a site, or starting an online chat—treat this as your goal and try to tie the online portion of the transaction to its offline outcome. At a minimum, provide a unique phone number on the website and track call volumes to that number.

Primary metrics: referring URLs; visitor motivation; traffic volume by segment

Secondary metric: inbound links from social tools

The second big question for transactional sites is how many visitors come from where. You get and pay for your business by understanding which search terms, campaigns, social networks, and referring sites have the highest conversion rates. You should also understand why visitors come to your site so you can be sure you’re satisfying their needs and putting appropriate offers in front of them. VOC surveys can reveal visitor motivations.

When looking at advertising, distinguish qualified from unqualified visitors. If you sell cars, young children who visit the site aren’t likely to be buyers and will reduce the ROI of your ad campaigns. Consequently, you need to eliminate unqualified leads from your conversion analysis and adjust your advertising spend to ensure that such leads don’t poison otherwise healthy conversion numbers.

Primary metrics: site effectiveness; ad and campaign effectiveness; usability

Secondary metrics: findability and search effectiveness; content popularity; community ranking and rewards

When you’re monitoring conversions and segmenting traffic, you need to focus on site effectiveness. Transactional sites thrive by upselling visitors. As more and more users buy through onsite search tools, it’s also important to monitor the effectiveness of searches and determine which search terms lead to purchases most often. The key to maximizing site effectiveness is constant experimentation in the form of A/B testing and multivariate analysis.

Since peer recommendations (“Other people who bought this also bought...”) are one of the strongest influences on purchase upselling, it’s also critical to monitor the user community for comments, ratings, and other feedback, and to borrow a page from collaborative sites by encouraging visitors to comment and recommend products.

Finally, you need to look at usability: examining how users interact with the site, particularly at places where abandonment occurs, is essential if you’re going to improve usability. Do users scroll down? Does your offer appear above the fold? Do buyers click on the button or the text? Does the page take a long time to load? All of these factors can impact how effectively you maximize each visit.

Primary metric: reach

Secondary metrics: enrollment; loyalty

Transactional site operators care about their ability to send messages to their users and have them act on those messages. Email campaigns, RSS feeds of promotions, and similar forms of enrollment are all useful, but they only count when someone clicks on the link.

You also want to know the lifetime value of a customer. You may find that a particular segment of the market purchases considerably more over months or years, and it’s wise to cater to that segment in your marketing, positioning, and offers. As retailers move away from broadcast and toward community marketing, enrollment and loyalty may overtake reach as the key measurement of relationship strength.

Primary metrics: impact of performance on outcomes; traffic spikes from marketing efforts; seasonal usage patterns; consistency of performance and availability

Secondary metrics: availability and performance; capacity and flash traffic

Transactional site operators care about performance primarily for the way in which it affects conversion. If performance degrades, conversion rates will fall. The site may be slow because of sudden spikes in traffic, peak seasonal periods (such as holiday shopping), content changes, or modifications to code and infrastructure.

Of particular interest is the impact of marketing campaigns on performance. Your marketing efforts must be tightly tied to capacity planning and performance monitoring to ensure that a successful marketing campaign doesn’t backfire and break your infrastructure.

Primary metric: how people are finding competitors

Secondary metrics: site popularity and ranking; relative site performance; competitor activity

If you’re running a transactional site, you care how people are finding your competitors, and how they can find you instead. This is often a battle of keywords and search terms, and in terms of organic search it is also a matter of how relevant Google and others think you are.

You also care whether you’re fast enough. It’s not necessary to be as fast as possible, but you should compare your performance to relevant industry indexes to ensure you’re not falling behind what users consider acceptable. On the Web, your competition may not be who you think it is. In addition to other sites that offer the same products and services, you’re competing against the expectations set by other websites your target market frequents. If those sites constantly improve and innovate while you don’t change, your audience will eventually grow disenchanted, even if those sites don’t compete with you in the traditional sense.

Secondary metrics: trolling and spamming; fraud, privacy, and account sharing

For transactional sites, most risks come from password and credit card leaks, which are a matter for your security team. But if you’re letting visitors rate and rank products, you need to be on the lookout for abusive behavior. For example, in an effort to improve the quality of ratings, Apple’s App Store chose to limit reviews to only those visitors who had purchased an application (www.alleyinsider.com/2008/9/apple-flexes-even-more-muscle-at-the-iphone-app-store-no-reviews-till-you-pay-up-aapl-).

You may also care about account sharing—if multiple users, all with unique tastes, share one account, your suggestions and recommendations will be less accurate and will undermine upselling attempts.

Secondary metric: site reputation

Your site’s reputation figures in both word of mouth and search engine ratings. But as a transactional site operator, you care mostly about which online conversations are leading to conversions, rather than reputation for its own sake.

Secondary metrics: API access and usage; mashups, stolen content, and illegal syndication; integration with legacy systems

If you’re running a transactional site, you care less about your content being used elsewhere, particularly if it helps spread the word about your products and your brand. Travel site Kayak.com, for example, compares flight prices across many airline portals, but it makes its money through affiliate fees from the airlines from which the visitors ultimately buy tickets.

On the other hand, if somebody is scraping pricing data from your site for price comparisons, it can undercut your margins and lead to price wars, so you need to identify hostile crawlers that harvest content from your site. You can then set up a robots.txt file to block well-behaved crawlers from those parts of the site, then identify those that ignore it and block them by user agent, source IP address, or a CAPTCHA test.

If your transactional site is a large-scale marketplace, you may have an ecosystem of buyers and sellers who’ve built tools around your application. There are hundreds of tools for eBay sellers, for example. You need to monitor these interactions so you don’t alienate power users, but also so they don’t violate terms of service.

If you’re selling online, you may also have backend connections to payment systems (such as PayPal) that need to be monitored as part of the overall site health.