In two previous posts, I explored the messy reality of Web site sprawl and looked at the big picture for getting things under control. The final piece of the puzzle before diving in to solving the problem is a basic strategy. If you’re a big company with a confusing plethora of Web sites, or a small company that wants to pre-emptively avoid sprawl in the first place, the steps below lay out a clear plan of attack.
Step 1: Audit your Web ecosystem
Before you do anything else, you need to understand the full extent of your potentially far-flung Web ecosystem, which means tracking down just how many sites are out there (along with who owns and manages them). This accounting will include at least primary domains and subdomains for US and international sites (e.g.,
acme.co.uk). It might also include sites within other sites that are functionally standalone (e.g.,
acme.com/jobs), which can often be tricky to identify. With the list of sites in hand, you can dive in to bringing things under control.
Step 2: Understand your audiences’ needs
Nature abhors a vacuum, and so does good experience design for Web sites. The first (and best) thing to help fill that vacuum is knowledge about your audience(s). You need to be able to answer the question, "What do people need (from all of my sites)?" If you don’t understand what people want (in terms of content or site functionality), then you have no basis to make decisions about what to do with all of those sites.
The best way to understand what your audiences want is to talk with them. Ideally this is accomplished through primary research (i.e., interviews with people that match your audience types). When primary research isn’t an option, or to just augment what you’ve learned, you can turn to secondary research. Sources of secondary information are things like customer feedback surveys, common requests or complaints to customer support, comments on your Facebook page, and mentions of your company and/or sites on Twitter.
Step 3: Figure out what’s broken
Now that you know what your audiences want out of your Web sites (or what they’re not getting), you can catalogue what’s broken. A few common things that can happen with Web sprawl are the following:
- Poor discoverability: People can’t find the stuff you have. It could be content that’s going to meet their primary needs, or other good stuff you have to offer that they didn’t even know they needed.
- Redundancy: You may have similar types of content spread over multiple sites, which can often lead to inaccurate or confusing information.
- Staleness: While this can be a problem for single Web sites, it’s even harder with lots of sites. When you’ve got a lot of content to manage and it’s spread all over the place, things fall through the cracks more easily, and content gets out of date. In the worst-case scenario, entire sites can be out of date or irrelevant (e.g., marketing campaign microsites with a limited lifespan).
- Frustration: Your audiences may get frustrated if they have to visit multiple sites to meet their needs. This can be especially vexing if your sites all have different designs and navigational systems; people are forced to figure out each site as they try to get something done.
In addition to the problems above (which are mostly challenges your audiences face), your site sprawl may also be creating internal issues (operational, technical or managerial). Do an internal assessment to identify all of these problems, too.
Step 4: Figure out what caused the sprawl in the first place
The Patterns section of my previous post identifies a few things that can lead to sprawl. Your company may have experienced these patterns or others, and it’s important to identify them. In some cases, the pattern / problem is fixable (e.g., lack of a centralized publishing platform). In other cases, it wasn’t really a problem that generated the sprawl in the first place (e.g., product or service differentiation), in which case there’s nothing to be done.
Step 5: Heal thyself where possible
If the fixable problems identified in Step 4 aren’t fixed, then in many cases it’s tough to address the sprawl they created (or if you did, the sprawl would just come back). So, address these problems if you can. If the problems are too tough to tackle, then come up with mitigations that keep them from generating sprawl in the future. One example of a mitigating approach would be to put in place a policy that any Web site that goes up without approval will be taken down (assuming you have centralized control of domains and DNS).
Step 6: Develop criteria for standalone sites
Based on the shape of your Web ecosystem, and a knowledge of what spawned the horde, you can develop criteria that help you decide whether an existing site should be standalone, or whether it’s best to consolidate with another site. One example is "Site is for a one-time event"; it makes sense in many cases to have these sites separate from your main site, and then take them down once the event is over. The great thing about having these kinds of criteria is that they not only help manage sprawl in the future, but also help figure out what to consolidate in the present.
Step 7: Consolidate
With the previous steps under your belt, you’re ready to tackle consolidation. It’s the trickiest, and most time-consuming, part of the process because it usually demands some experience design work. Many companies lack internal experience design teams, so they’ll often hire an outside vendor.
Consolidation requires a few steps:
- Appoint a single champion to oversee and manage the consolidation process. A strong leader will help manage what can be a complicated undertaking involving numerous stakeholders with differing needs.
- Figure out the target site for consolidation (i.e., what site will serve as your umbrella). It might already exist (e.g., your main company Web site), or you may need to design and build a completely new site, which is actually easier in some cases.
- Identify the sites that will be consolidated (into the target site) using the criteria identified in Step 7.
- Put together an experience design team (or hire an external vendor) to design the new, consolidated experience. At the very least, your consolidated site will need robust navigational systems to encompass new content and functionality. In addition, the new experience design should address your target audiences’ needs (from Step 2) and fix known usability problems (identified in Step 3).
- Design the consolidated site. Include the right internal stakeholders in the design review and approval process, balancing their feedback and input with the needs of your audiences and experience design best practices.
- Build and test the consolidated site.
- Launch the new site once the known bugs in the site fall below some agreed-upon threshold. Never launch a site with serious problems. If you do, prepare for a firestorm; you can bet the response to your buggy new site will be something like the healthcare.gov debacle.
- Shutter the sites you consolidated (and make sure you’ve got 301 redirects in place so people can find your content easily through search engines).
Step 8: Kill Zombie sites
The last step in the process is to clean up any of those straggler sites that were identified as Zombies (i.e., sites that are no longer relevant). You’ll probably want to set up 301 redirects for these sites as well, pointing people to relevant content in your new consolidated site (or at least to the home page, if it makes sense).
It’s not easy to manage sprawl, even if you have a solid strategy in place. It takes time, effort and money. The benefits to your business should, however, be clear:
- Improved usability and streamlined access to content means more satisfied customers
- More satisfied customers means reduced customer support costs (i.e., people can find what they need more easily, and don’t contact customer support as often)
- Streamlined content and Web publishing processes means less operational effort (and cost) to maintain your Web presence
If you’ve gone through the process of consolidation, or are just facing the challenge, I’d love to hear your thoughts.