Aside from acquiring better qualified visitors to the site, the next best strategy to increase average order value is getting customers too buy more items each time they purchase.

Definition
The average number of items per cart is the measurement of the number of units or items in each successfully completed cart:

Sum of Products Purchased / Number of Completed Shopping Carts = Average Items per Cart

To make this calculation the analytics package or commerce application need to be able to report on the number of items contained in each completed cart. If your particular application does not report on this value automatically, you may want to consider using a custom variable, being sure to only sum the number of products purchased for successfully completed carts.

Presentation
It is a good idea to present average items per cart along with average order value to provide context if one or the other KPI decreases.

Expectation
Depending on what they’re selling, retailers will quickly realize that average items per cart are usually very close to 1.0 and very difficult to increase. In some instances, this KPI is uninformative because it will always be a single item; in other instances this KPI can provide valuable insight into the disposition of visitors coming to the site and the quality of up-sell and cross-sell presentment.

Action
This KPI should be carefully watched when an effort is being made to improve the quality of up-sell and cross-sell functionality in the shopping cart. In situations where new strategies are being rolled out but average items per cart and average order value are unchanged, additional work is warranted.

In situations where this KPI suddenly decreases, it is worthwhile to review with marketing groups what changes if any have recently occurred. Perhaps a successful sale on a single item is underway and is decreasing the number of multiple-item carts being completed. The converse is also true; if no recent work has been done on how up-sell and cross-sell is presented, an increase in average items per cart may be indicative of a more qualified audience or a particularly successful campaign.

For retailers, average order value is considered a “key” key performance indicator by many, when combined with revenue per visitor or visit and order conversion rate, is essentially the pulse of the web site.

Definition
The basic calculation is:

Sum of Revenue Generated / Number of Orders Taken = Average Order Value

In the ongoing effort to optimize the online business there are two major KPIs describing the site’s ability to generate revenue: average order value and order conversion rate. Smart business owners work diligently to improve both but segmenting visitors and marketing campaigns into high, medium and low AOV groups can help identify where the “best” (e.g., high AOV) customers are coming from.

Presentation
As with other dollar-based KPIs, presentation should be fairly obvious. It is a good idea to present this indicator and average cost per conversion, order conversion rate and revenue per visitor together to provide context to each.

Expectation
Sites should determine a baseline AOV for all customers to use as a comparator for all marketing acquisition campaigns. For example, it might help to make and keep track of the average order value for the entire site, targeted email campaigns, untargeted email campaigns, search marketing efforts and so on. Assuming your conversion rate is same for all customer acquisition efforts (rarely the case), you’ll discover that you’re better off focusing your efforts on high-AOV generating campaign types.

Entire Site AOV Email AOV Keyword AOV Banner Ad AOV
$100.10 $95.50 $120.15 $101.25

As you can see, the average order value for customers associated with search keywords is 20 percent higher than the site-wide AOV.

Action
A decrease in average order value should be compared to changes in the order conversion rate. If AOV decreases but order conversion rate increases revenue per visitor should stay roughly the same; if AOV and order conversion rate both drop revenue per visitor will likely be strongly impacted. Regardless, average order value should be closely watched and any changes should be diagnosed, looking at changes in the checkout process and marketing acquisition programs.

This key performance indicator makes the list of “RED BUTTON” KPIs that, when they go wrong, should bring everyone to a screeching halt while the problem is diagnosed. Especially when compared to marketing acquisition indicators like average cost per visit, the value of conversions are critical.

Average revenue per visit is a more granular examination of your site’s financial performance but otherwise similar to average revenue per visitor.

Definition
See average revenue per visitor but substitute “Visits” for “Visitors.”

Presentation
See average revenue per visitor.

Expectation
While average revenue per visitor is really a long-term, time independent performance indicator, revenue per visit is a good indicator of how you’re doing right now in your marketing and conversion efforts. Compare revenue per visit to average revenue per visitor to see if your short-term efforts are paying off but not really contributing to the lifetime value of a visitor.

Action
See average revenue per visitor.

(I listened to your comments and appreciate the feedback. What do you think?!)

Often it pays dividends to keep track of the cost of driving individual visits to the web site for comparison to your average cost per visitor. These key performance indicators used in tandem can tell you a great deal about your marketing acquisition costs.

Definition
A function of the total sum of marketing costs, the average cost per visit is defined as:

Sum of Acquisition Marketing Costs / Visits = Average Cost per Visit

Challenges associate with calculating this key performance indicator are the same as average cost per visitor.

Presentation
It is a good idea to present average cost per visit and average cost per visitor side-by-side, depending on how different these calculations are.

Expectation
In an idea world you would be able to drive visits with little or no marketing costs; unfortunately it is far from an ideal world. Still, lower is better.

Action
Especially when experimenting with new marketing channels you want to watch your average cost per visit carefully, looking for a dramatic increase that is not correlated with increases in value-based KPIs like average value per conversion, revenue per visit or average order value.

Francois Lane sent me this idea for a key performance indicator:

Here it is:Price of the Product x Conversion Rate of the Page = Average Revenue per Page View

This metric is interesting for e-commerce with large offering, like Amazon for example. It gives, in dollars, the revenue generated (on average) each time the product page is displayed. Then, a way to increase revenue of the store could be to promote on frontpage the highest RPPV products, then on product category pages, etc.

Or you could replace the price value with the “gross profit” of the product to maximise profit instead of revenue.

I love it! This kind of indicator would be slotted into the special category of “list” KPIs and would likely sort out the top ten revenue generating pages, allowing retailers to watch for gainers and losers.

Excellent suggestion, Francois! Are any of the rest of you reading this blog using this kind of KPI? If so, how has it been working for you?

A reason to describe KPIs for “Average Revenue per VISITOR” and “Average Revenue per VISIT”? Same problem for cost per visitor and cost per visit?

For some reason I cannot think of a good business reason that a company would want to differentiate these two metrics.

Can any of you? If so, I’d love to know!

Revenue per visitor is a critical metric but not just for online retailers and advertising supported sites. Marketing sites can better understand their marketing efforts by estimating value based on conversion events and customer support sites can approximate revenue supported.

Definition
In general:

Sum of Revenue Generated / Visitors = Average Revenue per Visitor

Each business model will calculate revenue generated or supported differently:
• For retail sites the sum of revenue generated is easily calculated.
• Advertising-based sites can use the sum of advertising revenues generated or a calculation of average CPM times impressions served.
• Marketing sites focused on lead generation are encouraged to estimate the value of leads generated by comparing similar quality leads to past results.
• Customer support sites should ideally sum the amount of customer contract value supported by the site. For example, if you know that 100 people are getting support for a $100 product and 50 people are getting support for a $500 product, the sum of revenue supported would be 100 x $100 + 50 x 500 = $1,250,00

While the customer support case is obviously artificial it serves no less value for sites to track the value of visitors they support.

Presentation
As with other dollar-based KPIs, presentation should be fairly obvious. The only exception would be for the customer support model in which the indicator should be clearly titled “Average Revenue Supported per Visitor.”

Expectation
As you would expect, the more revenue per visitor you’re able to get, the better off you are. The obvious strategy for improving this performance indicator is to attract more valuable visitors to your web site. Consider using average revenue per visitor to critically examine each new visitor acquisition effort, segmenting as necessary, to determine whether different strategies are actually working.

Action
If this number drops off suddenly or precipitously likely the first call you should make is to your marketing department and the next to your operations group. Often times either a large group of unqualified visitors has been attracted to the site or something has gone wrong with your revenue realization path (e.g., your shopping cart is broken or your site is performing slowly, thusly reducing the number of advertising impressions you serve.)

This key performance indicator makes the list of “RED BUTTON” KPIs that, when they go wrong, should bring everyone to a screeching halt while the problem is diagnosed.

Regardless of your business model, conversion is one of the most important visitor activities you need to track. By calculating the average cost per conversion you can ensure that you’re not paying too much to acquire visitors.

Definition
The general calculation for average cost per conversion is similar to average cost per visitor and average cost per visit:

Sum of Acquisition Marketing Costs / Conversions = Average Cost per Conversion

Sophisticated marketers may want to segment this KPI for individual conversion events; to do this you need to have a pretty good system for tracking marketing costs so that they may be associated with the intended act of conversion. For example, if your site is designed to generate leads but visitors can also sign up for a newsletter, you may want to assign the lion’s share of marketing costs to the former and a small fraction to the latter—only the marketing you do to grow your newsletter subscription base. Doing so will inevitably produce a better-looking KPI for your newsletter subscription conversion event but this makes sense as long as the latter event is ancillary to your marketing goals.

Similar to average cost per visitor it does make sense to segment average cost per conversion by marketing channel to help identify strategies that are ineffective from a cost perspective.

Presentation
Because this KPI is dollar-based it and critical to the success of most businesses it is unlikely you’ll need to change much in the presentation. It is worthwhile, if you break down your cost by conversion event, to both provide a global view (all marketing costs divided by all conversion events) for reference and also clearly identify the conversion event for micro-events.

Expectation
If you’re paying more for conversions than the conversions are worth then clearly something has gone wrong. For most companies this is not the case and the expectation is that even nominal ongoing savings in conversion costs can add up. By constantly re-examining your marketing acquisition efforts and cutting waste, your cost per conversion can be dramatically improved.

Action
Any time average cost per conversion increases it is advised to immediately examine your marketing efforts to see what has changed. The most common case is that some expensive program has recently been launched and is failing to drive an appropriate number of conversion events. In this case you usually don’t want to immediately cease the marketing activity in question but do want to pay close attention to said effort, watching for any improvement.

Visitor acquisition costs often spiral out of control when left untracked. While tracking these costs can be difficult in the long run the effort is worth it.

Definition
A function of the total sum of marketing costs, the average cost per visitor is defined as:

Total Acquisition Marketing Costs / Visitors = Average Cost per Visitor

For most companies the tricky piece is summing acquisition marketing costs, owing to the fact that few companies are accurately tracking these numbers on anything more granular than a quarterly basis. It is recommended that you limit the summation to online marketing activities only unless you strongly brand your URL in offline marketing materials. By adding up the costs of search, email, banner, partner and feed-based marketing activities a fairly useful KPI can be generated.
This indicator is a good candidate for segmentation by marketing channel. For example, you may want to calculate the average cost per visitor for your email, banner and search based marketing efforts.

Presentation
Because this KPI is dollar-based little usually needs to be done regarding presentation to attract stakeholder interest. Especially if the average cost per visitor is high, most executives and managers will pay close attention to this indicator.

Expectation
Ideally visitor acquisition costs are low and contribute to a well-run, high margin business. Unfortunately the ideal case is rarely observed. It is worthwhile to set the expectation that the company will work diligently to lower visitor acquisition costs and carefully critique each marketing channel.

Action
If cost per visitor suddenly increases it is worthwhile to compare this increased cost to average revenue per visitor and relevant conversion rates. If cost per visitor is going up but revenue or conversion are flat or decreasing something has gone awry. The converse is also true: if your acquisition costs drop suddenly you want to make sure that this fortuitous event has not happened at the expense of revenue or other measured value.

Most companies forget to track one of the most important customer support metrics there is: the amount of time it takes you to respond to a customer request sent via email.

DefinitionThe average response time for an email inquiry is a measurement of the number of minutes, hours or days it takes you to provide a visitor a human-generated response to a email-based inquiry:

Sum of Response Times in [TIME UNIT]/ Total Number of Email Inquries = Average Time to Respond

The [TIME UNIT] in this equation refers to minutes, hours or days, e.g., “Sum of Response Times in Days”. Response time is defined as the difference in [TIME UNITS] between the time the inquiry is received and the time that someone in your company answers the email. While there are a handful of technologies designed to automate responses, rare is the substitute for a personal email responding to the question or concern. Any company concerned with how visitors perceive their commitment to customer suport is advised to respond personally to these inquires.
While summing these times can be arduous, the process can be simplified by creating a central spreadsheet of inquiries and responses or mining your customer support application for the data.

Presentation
Because nothing is more frustrating to visitors than sending an email and having to wait endlessly for the response, this KPI is one that lends itself well to conservative alerts and warnings being generated. Depending on your particular business, you should set the warning threshold very low and use warning generation as a strong action driver.

Expectation
Your visitors and customers expect a near-instantaneous response to any email they send you, especially when they have a problem. If you want happy customers and prospects you should consider setting expectations of response times very low, e.g., less than 6 hours or under one day—same day response. As an exercise, track this KPI against the volume of calls into your organization to see if a 10 percent improvement in average response time correlates well to a 10 percent decrease in call volume. You may be pleasantly surprised.

Action
Regardless of your average response time this KPI should never get worse and increase. Any sustained increase should immediately be investigated, looking to see if perhaps there has been an increase in complex inquiries, an extended illness or problem among those responsible for responding or worse, someone completely ignoring requests for help.

Average visits per visitor over a finite timeframe can help you understand how much interest or momentum the “average” visitor has.

Definition
The total number of visits divided by the total number of visitors during the same timeframe.

Visits / Visitors = Average Visits per Visitor

Sophisticated users may also want to calculate average visits per visitor for different visitor segments. This can be especially valuable when examining the activity of new and returning visitors or, for online retailers, customers and non-customers.

Presentation
The challenge with presenting average visits per visitor is that you need to examine an appropriate timeframe for this KPI to make sense. Depending on your business model it may be daily or it may be annually: Search engines like Google or Yahoo can easily justify examining this average on a daily, weekly and monthly basis. Marketing sites that support very long sales cycles waste their time with any greater granularity than monthly.

Consider changing the name of the indicator when you present it to reflect the timeframe under examination, e.g, “Average Daily Visits per Visitor” or “Average Monthly Visits per Visitor”.

Expectation
Expectations for average visits per visitor vary widely by business model.

• Retail sites selling high-consideration items will ideally have a low average number of visits indicating low barriers to purchase; those sites selling low consideration items will ideally have a high average number of visits, ideally indicating numerous repeat purchases. Online retailers are advised to segment this KPI by customers and non-customers as well as new versus returning visitors regardless of customer status.
• Advertising and marketing sites will ideally have high average visits per visitor, a strong indication of loyalty and interest.
• Customer support sites will ideally have a low average visits per visitor, suggesting either high satisfaction with the products being supported or easy resolution of problems. Support sites having high frequency of visit per visitor should closely examine average page views per visit, average time spent on site and call center volumes, especially if the KPI is increasing (e.g., getting worse.)

Action
All web sites desire some kind of relationship with their visitors over time—the wild cards are usually the type of relationship and the amount of time. Customer support sites want people to visit whenever they have a problem but don’t want customers to have problems per se yielding a high average visits per visitor over a longer period of time. Retail, marketing and advertising sites all want people to come back all the time to buy, learn or click respectively. The challenge for site operators is figuring out how exactly to drive this return activity and knowing what to do when it fails to appear.

For the most part, when this KPI trends in the wrong direction you need to ask “what just happened?” Your average visits per visitor should be relatively stable providing your site has been available for at least 6 months and you’ve not made any major changes to the site or your retention marketing strategy. Therein lies the opportunity: If you change your retention marketing strategy or your site you should expect to see a change (albeit slight) in this KPI in the following weeks and months. If none appears, what went wrong? If the KPI improves dramatically, great! Understand what you did well and repeat as often as possible.

If this KPI suddenly gets worse, figure out why. Common culprits include site changes breaking bookmarked links, the emergence of a new competitor and the intangible offline “vibe”, e.g., perhaps you’re just no longer as cool as you think. Keep in mind before you panic: You need to give your visitors enough time to return and visit depending on your business model.

Average page views per visit are an excellent indicator of how compelling and easily navigated your content is.

Definition
The total number of page views divided by the total number of visits during the same timeframe.

Page Views / Visits = Average Page Views per Visit

Sophisticated users may also want to calculate average page views per visit for different visitor segments.

Presentation
Presentation of average page views per visit can be supplemented by associating the monetary value of a page view for the advertising business model. Based on an average cost per thousand (CPM) advertising impressions, you can calculate the value of the average visit as follows:

Average Dollar Value / 1,000 Page Views * Page Views / Visit = Value of Average Visit

For example, an advertising site having an average CPM of $25.00 and an average 3 page views per visit would make the following calculation:

$25 / 1,000 page views * 3.00 page views / visit = $0.075 per visit

Expectation
Expectations about average page views per visit depend on your business model.

• CPM-based business models that depend on high page view volumes should work to increase the average number of page views per visit, thusly increasing the value of each visit.
• Marketing and retail sites generally want to increase this average, indicating a greater interest on the part of the visitor. However, depending on the specific goals of the site, more page views can indicate confusion on the part of the visitor.
• Customer support sites generally want to decrease the number of page views per visit, at least in sections specifically designed to help visitors find information quickly.

Action
When the average number of page views per visit trend against expectations, I recommend examining a handful of common site components that affect page views:

• Navigational elements (e.g., your information architecture). If it is difficult for visitors to navigate your site they will often be forced to view more pages as they hunt. Conversely, if your site is difficult to navigate, visitors may leave your site prematurely out of frustration.
• Content. If your content is poorly written and doesn’t follow best practices for writing for the web, visitors may leave your site prematurely. Conversely, if your content is well written, visitors may be inspired to “keep reading”, driving up the average number of page views.
• Search technology. If your search functionality is poor, visitors may be forced to click to look for information. Conversely, if your search functionality is good, visitors may be leveraging search, thusly reducing the number of pages viewed.
• Marketing efforts. If your marketing efforts are poorly targeted, visitors are less likely to view many pages. Conversely, if your marketing efforts are good, visitors may view a large number of pages.

It is worthwhile to use the KPIs for average time spent on site and average time spent on pages for key pages when diagnosing problems with average page views per visit. You may also want to look at how your internal search application is being used by examining percent visitors using search, percent “zero result” searches and average searches per visit.