October 17, 2011
As I mentioned last week, a good website testing plan answers four basic questions: Who? What? When? Why? But a common folly in website testing is that marketers approach these questions sequentially rather than holistically--selecting targets that are too narrow and, as a result, producing tests that take too long to achieve statistical significance. In this second post in the "Strategies for Successful Testing" series, we'll examine five segments that make for great answers to this first question (Who?), and offer some initial thoughts on what can you show them to start developing the ingredients of a great test campaign.
5 Key Segments You Should Target
1) New vs. Returning Visitors: New and returning visitors engage your website in entirely different ways, and your test campaigns should do the same. Returning visitors have a certain familiarity with your company brand, site structure, and navigation (and all things equal, you're more likely to convert them). By contrast, you should assume that new visitors know nothing about you at all.
New and returning visitors have the added benefit of encompassing 100% of your site traffic. Regardless of a visitor's traffic source, landing page, or geography--everyone is either 'new' or 'not new', so tests that are split along this line typically achieve statistical significance sooner than any other segment. A few very simple tactics can make the site experience more relevant to each of these groups. For example, you can serve a "Welcome Back!" message to visitors you've already cookied and -- for new traffic -- move your "Get Offers in Your Email" functionality to the top of your page template to drive more sign-ups among first time visitors.
2) Branded vs. Nonbranded Search: Like new vs. returning visitors, your traffic from branded and nonbranded queries will engage your site in fundamentally different ways. A visitor searching your brand was hoping (and expecting) to find you. By contrast, visitors searching nonbranded queries were expecting not to (necessarily) see you in the search results.
Oftentimes, nonbranded searches can be difficult to convert because these prospects tend to be in "window shopping" mode and seeing how you stack up to the competition. But try this on for size--converting nonbranded traffic is a lot easier than you think (perhaps even easier than converting your brand). The challenge with branded searches is that you rarely know -- with any specificity -- what the visitors actually want. Sure, they wanted you, but which product or product category? With nonbrand, knowing that a visitor searched for a "50-inch LCD TV" or "Stainless Steel Blender" can go a long way toward helping you tailor the shopping experience to satisfy those needs.
A great way to increase your nonbranded conversion rate is to echo your ad text on the landing page (and throughout your website). This establishes messaging consistency and confirms for the visitor that they've landed in a relevant place. For branded visitors, you may find that most are also 'returning', so you can employ the same tactic above, but with a more specific message such as "Welcome AdWords Shoppers!"
3) Abandoned Cart: You came so close! You attracted prospects to your site, got them to the right page, and convinced them to place a product in the shopping cart. Then something went wrong! You may not know why any particular visitor fails to complete a purchase, but you don't have to take it sitting down. To be sure, you should strive to reduce cart abandonment as much as possible, but consider a banner or light box reading "You Still Have Items in Your Cart" to remind returning visitors that they can complete their purchase in a few short minutes. And if you know their email address, consider sending a reminder email when an in-cart product remains unpurchased after 24 hours. A great email will include a description of the product's features and links to both the shopping cart and any product reviews available.
4) Days Since Last Purchase: You've already done the hard part by convincing these visitors to purchase from you, and now they're back on your doorstep. Keeping customers is usually easier than winning new ones, but don't let this make you complacent. If you have data that shows, on average, how long your product lasts (think coffee or shampoo), then consider serving replenishment offers to consumers who return after a certain number of days. Step it up a notch with a coupon code that's automatically applied when the visitor goes to Checkout.
A Cautionary Note: Returning Consumers are just a fraction of your overall site traffic, so you're entering a territory where it can be very difficult to achieve significance within 2 weeks. This is fine--as long as you start big (e.g., New vs. Returning) and work your way down with the insights learned along the way. With segments this small, think of your efforts less as "tests" and more as just common sense targeting.
5) Current Weather Conditions: Whether it's hot and sunny, or cold and snowing, weather changes buying behavior in significant ways. Your brick-and-mortar counterparts routinely change store fronts based on current weather conditions--and so you should change your website! You can adjust the products featured on your Homepage -- and the order in which they're rotated -- based on whatever Mother Nature's up to. For example, you could promote jackets, scarves, and gloves every time the temperature dips below a certain point, or push sunscreen products when it's both sunny and warm.
Regardless of the segments you choose to target, be sure to consider the four questions above together. Your objective is to build a great test campaign--not simply to identify the best audiences. Next week, we'll dive more into the ways you can improve conversion rate for each of your audience segments. Stay tuned for "Strategies for Successful Testing: Segmenting Your Way to Great Results."