You need to see over the walls that separate your websites.
What’s that mean? Imagine you own a casino; you know a player lost $100 at, say, craps, but you want to know where each patron came from, how long each stays in the casino, and how each moves from the ATM to the craps table to the bar to the ATM (my typical path).
This is like a website (let’s call it fake.com) with a number of peer content sections, each served on different sub-domains (community.fake.com, support.fake.com, and www.fake.com). In some web analytics tools, a visitor that surfs from Community to Support is counted as two distinct visits: one visit from the true external traffic source (say, organic search) but showing a faulty exit, another visit mis-attributed to a referring site (community.fake.com), and both visits inaccurately seen as less engaged than they really are.
On the other hand, imagine you own a grocery store in a tough part of town, so you sit at the register behind smudgy bulletproof glass (stick with me, here); you know a customer bought, say, pepper spray, but you want to know how each customer found the store, which aisles each walked, and which displays each passed before checkout.
This is like a website where purchases are completed on a top-level domain (fake.cartvendor.com) or a sub-domain (cart.fake.com) separate from the main website. In some web analytics tools, all the revenue is counted as coming from the referring site “fake.com” instead of being attributed to the true external traffic source that brought people to the brand in the first place: PPC, email, banner ads, affiliates, etc. Revenue can’t be attributed to promotions visible on the main domain, either.
Providing marketers with a complete view of these “cross-domain” visits in a tool like Google Analytics requires customization of the tracking code on every page and often of the links and forms that take a user from one domain to another. Such a universal change to tracking code in turn requires a thorough quality-assurance effort to make sure visitor behavior is captured as desired, whether you have a non-public staging environment or not.
Transitioning from a separate, walled-in view of each domain to a holistic view of the whole multi-site empire takes some planning. Which accounts or profiles will retain historical information and will they be apples-to-apples with reporting on visits going forward? How do we name accounts, profiles, and even website pages (which can be re-named in reporting for clarity) so the analyst of the future knows which data is where? These decisions are especially tough when mergers and acquisitions bring together multiple domains.
Some stains from the old domain-by-domain reporting are hard to remove. For instance, so-called “self-referrals” – when you see your site’s name in the list of external sites that send traffic – muddied traffic source reporting before the cross-domain upgrade…and they will still show up in reporting afterwards, thanks to returning visitors who were cookied with that erroneous traffic source in the bad old days.
Whatever type of empire, gaining the perspective to watch visitors as they move among your different domains is worth the investment. For e-commerce businesses, which often use shopping carts hosted elsewhere, attributing revenue to the correct marketing channel better allocates the marketing budget. For informational websites, knowing how visitors really navigate the site – including true entry and exit points – enables data-driven decisions about content, audiences, and user experience.