Leading Indicators Of The Future

The web is always changing. Growth in web usage has led to growing traffic, and sometimes growing revenues. But, high growth in overall use of the web can hide important information about changes in how visitors feel about a particular site — changes that could lead to problems in the future.

Say a web site grows at 10 percent per month. By plowing growing profits (whoops, make that revenues) back into promotion, revenue growth can be continued, right?

Maybe so, but it’s also possible that the current growth is hiding important information about the future, which can be extracted from existing profile and traffic data.

There have been many times when the growth rate for a technology or a sector of the economy has turned dramatically in a short period of time and caught everyone off guard. Well, practically everyone.

Way back in Economics 101, professors presented the concept of “leading indicators” to foretell the future. Economists have identified the 12 leading indicators of the US economy.

For an e-commerce web site, it doesn’t take a statistician to predict that less traffic means there will soon be less revenue. But, there is a lot more to forecasting web activity and revenue than just looking at traffic reports.

Forecasting, whether it’s the weather or revenue, requires the right tools, techniques, and experience in using both. For instance, web marketers with visitor profile information can find leading indicators hiding in their data — data that can help predict the future.

Brand managers have used frequent surveys of awareness as one indicator of future revenue. When awareness goes up, an increase in revenue can be expected to follow. Since there is a time lag between the change in awareness and the change in revenue, time-series analysis is one of the analytical tools used to measure how one affects the other.

Internet promotions have a very quick response, especially email promotions where click-throughs typically occur within a few hours to a few days of the initial mailing.

However fast people are to click over to a web page for product information, they are much slower about making a purchase decision. And, it’s that delay in making a purchase that makes it difficult to determine which marketing communications activity actually resulted in a sale.

When people use the web to gather information, they need time to read and digest that information before they actually decide to make a purchase. Thus, additional time passes before they actually begin shopping.

So, even though a click-through occurs within a few hours of sending a promotional email or running a banner campaign, it takes time for the consumer to actually pull out the plastic and place the order.

If you use “closed-loop” online marketing techniques, then you probably have the data needed to measure the lag time from promotion to purchase. Once the typical lag time is determined for a product line, it becomes easy to make projections about what to expect from future promotions.

Of course, very few projections are 100 percent accurate, but having a technique to measure the lag time and expected results creates a framework to evaluate individual promotions. Comparisons of promotions that performed better than projected with those that didn’t meet expectations can provide insight into which factors to adjust on upcoming marketing campaigns.

Statisticians and economists use a variety of techniques to adjust for the time lag between data as they try to identify reliable leading indicators. Many data mining products include time-series analysis, dynamic regression models, and other complex techniques.

But you don’t have to be so formal about analyzing data to find leading indicators and make useful forecasts. By combining the visitor profile with a history of each person’s web site activity, many measurements can be created that can give early insight into changes in growth, such as:

  • Number of visits per sale

  • Number of days between visits
  • Number of visits to the site per sale
  • Number of return visits to the site
  • Number of visits for customers versus non-customers
  • Number of pages visited for customers vs. non-customers
  • Time spent on the site for customers vs. non-customers

Each of these metrics can provide an interesting snapshot of activity during a month, but the real value is in tracking this data over time. For instance, if the “average number of visits to the site” compared to the “number of sales made” grows each month, it could indicate future revenue problems even if total site traffic is increasing.

By tracking performance measurements such as these each month, it is possible to develop an “early warning system” that can provide just enough lead time to change the promotion, product mix, or content to head off unwanted changes in revenue growth.

Related reading

women-in-tech
nurcin-erdogan-loeffler_wikipedia-definition-the-future_featured-image
pwc_experience-centre_hong-kong_featured-image
12919894_10154847711668475_3893080213398294388_n
<