|
|
Getting Sales From Forms
By Gary Angel
Expert Author
Article Date: 2008-02-01
In the first installment of this series on Forms, I talked about the critical role of usability testing in the development of good Form processes on the web.
Even after a Forms process roles out, a lot of the focus of analysis tends to be on its basic operational performance. Is a Form Page broken in some instances (usually because of untested field edits)? How long does it take users to traverse the Form? What error messages get triggered during Form completion? These are all important questions and they tend to get a lot of focus.
But one of the key real-world aspects of Forms behavior - and something that you absolutely CANNOT test in usability labs - is whether the Forms are doing a good job selling the product.
Forms aren't supposed to sell the product, are they? As usual, I'm going to answer, "It depends."
There is a significant school of design thought that emphasizes paring away every aspect of the Form that isn't absolutely necessary to get the job done. It seems logical. An almost inevitable consequence of some of the principles I elaborated earlier: that every Form element introduces at least some small amount of friction. Reduce friction and you improve Form performance.
This focus on operational perfection, however, isn't always ideal. Particularly for longer, more complex Forms, the issue is NOT clear cut. As friction builds up during the process, visitors who aren't totally committed to the purchase may drop off. Adding Form elements that reinforce or build the commitment to the sale may have a positive impact on the process despite the addition of some Form friction.
You can't test this in usability. Testers aren't actually buying the product. So no amount of usability testing can ever tell you how much - or whether -additional sales reinforcement during the Forms process is necessary.
Are there any behavioral cues that can help?
There are. But this is also a place where testing will likely be the ultimate source of answers.
If you're looking to make a case for testing (which can be expensive in Forms processes) or just to decide if sales reinforcement might be in order, then you should focus on a measure we call directional abandonment.
When users exit a Forms process, their manner of exit can be very informative. There are really three types of Form Exit: exited the site, exited to the rest of the site but did nothing of interest (returned to home page and exited or backed-out to the page prior to Form Start and exited), and existed to the rest of the site and looked at something.
This last category of exit is what we call directional abandonment. And it's extremely informative about what types of issues or concerns may be arising during the moments before a sale.
Don't expect Directional Abandonment to be an enormous percentage of exits. For most sites, process exits are mostly site exits. But even modest levels of directional abandonment can be very significant and can provide cues about the psychology underlying the much larger number of site exits.
There are different ways to look at Directional Abandonment. It's a web analytics task that is complicated by the inability of pretty much all WA tools to isolate the behavior of visitors AFTER an event.
In most cases, you'll begin with a Next Pages analysis. I believe this is most fruitful on a step by step basis. Different points within the Form will often yield different types of Exit behaviors. If you want to be fancy, start by building segments of visitors based on their deepest penetration into the Form. So if you have a five-step process, build a segment of visitors who reached Step 4 but not Step 5; then a segment of visitors who reached Step 3 but not Step 4; and so on. Building segments in this fashion let's you compensate for the fact that visitors will often slide backward in the Form before exiting.
Using these segments, look at Next Pages from each step of the Form. Throw out site exits and Form Pages, and you have a list of destinations. I usually classify these by Functional Types. So I'd look at how visitors left for Engager pages (like the Home Page), Router Pages (by subject), and Convincer Pages. Within the Convincer Pages, I'd want to look at how Exit To rates compared to Viewed and Entered From rates. I'm looking to find a pattern that suggests a specific Page(s) represents the area of primary concern.
Don't forget to always look at Search Keyword behaviors as well. As I've written about before, Internal Search keywords can be a powerful tool for understanding what visitors are looking for or care about at any specific point in a site visit.
Of course, Next Pages doesn't capture a significant chunk of interesting behavior for Forms exits. It's commonly the case that navigation outside the Form is limited to a few pages (such as the Home Page). So you need to be able to do path analysis from each page of the Form Process based on the "depth of penetration" segments.
In some other types of analysis, I'd suggest that you use an affinity measure. But that won't work here. Affinity analysis (viewed in same session) has no conception of sequence. And it's far too likely that visitors will have viewed the Convincer Pages BEFORE entering the Form for this analysis to be meaningful.
So you'll have to use pathing to establish how many times visitors exited each Form Step and went to any Convincer Page and the rate at which they subsequently viewed each specific Convincer page. By doing this, you'll often be able to isolate the page(s) that are most of interest to Form abandoners.
You may be wondering why I don't suggest just asking visitors what's wrong. It's not a bad strategy. But neither is it a replacement for behavioral analysis. This is a notoriously tricky place to capture good data. And even if you can get enough visitors to respond, it's difficult to get the right questions in place to really understand the behavior. It's certainly not a bad idea to ask "Why did you choose not to buy?" But don't expect the visitor to answer intelligibly. Buyers often won't tell you - "I needed to be persuaded about the value one last time."
How much (if any) reinforcement is necessary is not a question that can ever be settled in the abstract. It doesn't even have a single right answer for a site process - since changing the mix of visitors into a process can change the optimal process design.
It has been my experience, though, that many longer and more complicated forms processes can benefit from sales reinforcement. Done well, the frictional cost of reminding the user why going ahead is worthwhile can be very minimal.
I've learned through bitter experience that in the consulting business, it's never a good idea to assume the job is yours until the paperwork is actually signed. Forgetting to reinforce the sale (if there's an opportunity) during the "paperwork" stage has cost us more than one sale. The same is true on the web.
I would say that you can take it easy once you've got the sale, but that isn't really true either. Because in my next post I'm going to re-visit a topic I've talked about several times in the past: how to think about the last step in a process - the Thank You page.
Comments
About the Author:
Gary Angel is the author of the "SEMAngel blog - Web Analytics and Search Engine Marketing practices and perspectives from a 10-year experienced guru.
|
|
