“Under no circumstances will we change the basic layout of those pages!”
I heard this quote recently when discussing website changes for a major on/offline marketer. His point - and it was a legitimate one - was that significant changes would not only undo years of evolution, but may also confuse the customer. So we went with small changes - at least at first.
But as the first test results came in, the doors began to open on the possibility of making major changes.
So, is it legitimate to make wholesale changes to layout? To navigation? To page sequences? Yes.
If the "settled opinion" of the marketing group or the CEO on major elements of the site - layout of the homepage, for example - can be unsettled by the actual, measurable results of real customers experiencing the site, shouldn't those results be a major endorsement for change?
An existing website may have been shaped by usability studies, research, and contact with actual customers. But the volume of customers and the objectivity of live testing can provide insights that cannot be reached with a focus group. Think small and you may just lose this advantage.
Yes, there are hard-to-quantify impacts in marketing. A promotion that increases sales (like GM's employee discount) may damage the brand. If those are your concerns, use these guidelines to mitigate risk:
1. If the change fundamentally alters navigation, consider targeting the test only to new visitors at first.
2. Big changes usually impact many pages on the site. Make certain your testing tool can handle controlling content everywhere it appears so you never look inconsistent.
3. Offer the option to get back to the old look and feel.
There is always a better version that satisfies customers so that they will sign-up. If the "settled opinion" of the marketing group or the CEO on major elements can be unsettled by the real results of customers on the site, then we must pay attention.