Research is at the heart of UX. Without solid data to build upon, it's impossible to build truly functional and user-friendly experiences.
There are ways to go about gathering data without directly interviewing the users involved:
- A/B Testing
- Automated Remote Screen Recording
Let's take a quick look at how these might be effective in your case.
Gathering and drawing conclusions from analytics is a key skill for UX engineers and researchers. UX Booth has a great introduction into how to use analytics in a UX capactity and is definitely worth a read. In general, however, you'll want to do at least a couple of things to use
analytics to your advantage.
Setup an Analytics Gathering Tool
This may be more difficult given the government nature of your project, as there are some security concerns and most analytics tools require open access to at least their APIs. Still, there are many out-of-the-box options that require very little setup to be effective from day 1, of which Google Analytics is the most prominent and commonly used. In a pinch, it's possible to setup some basic analytics gathering using some server-side coding and a database, though anything you build would have a hard time being as fully functional as a dedicated suite like Google.
Set Key Performance Indicators
Key Performance Indicators (KPIs) are the metrics that you'll use to measure success. KPIs are different for every organization, and you'll want to spend some time nailing down what a successful re-design will mean to you and your client. While your KPI could be just about anything, there's one thing it must be: measurable. cxpartners.co.uk have a great write-up on common and uncommon KPIs that's worth a glance over as you begin to brainstorm what measurements will mean success in your project.
A/B testing is a method of testing that relies on presenting two different site designs to different users to test which is most successful. This can be done through moderated testing (not an option in your case), surveys (also not an option), or automated testing efforts in which the site randomly or semi-randomly assigns a site design to users when they visit. This requires additional software installation on the server, and therefor there are security considerations that need to be taken into account, especially in a government context.
Due to the constraints of documentation and the need to provide at least a moderately consistent experience to users, A/B testing is, by its nature, a rather slow way to evolve a website, as only one change can be effectively tested at a time. It also relies on analytics gathering to be successful, especially if done by automated site design assignment. Invisionapp.com has a great getting-started guide that you should read over for more informaiton.
Automated Remote Screen Recording
Automated remote screen recording is the screen capture of users on a site using a piece of software installed on the server (which, as we said before, comes with its own set of security concerns). This allows you to remotely review users interacting with a site at your leisure.
There are some great benefits here. For starters, you get to actually see users interacting with the site, and have a good idea of where they are stumbling or stalling or failing. It gives you a great idea of how users go about error recovery. What's more, you can track mouse movement, which some studies show has about an 84%-88% correlation to eye movement. From this data, it's possible to do create of the the traditional eye tracking metrics and data outputs with much less invasion and cost.
Check out this list from creativebloq.com for a good rundown of some of the available tools.
A Word of Advice
Here's the thing - setting up any of this will be far more expensive and time consuming than taking the time to actually interview your users. Yes, they may be busy people, but your client's reluctance to allow interviews is going to have a very real impact on your project's success, budget, and length.
Consider presenting these alternatives and also suggesting asking for a timeboxed, half-hour interview with each stakeholder. If you design your questions well, you shouldn't need more than that to nail down the basic usability concerns of your users. Most users are receptive to UX testing requests, as the testing shows that your organization is engaged and interested in helping solve their problems. In the end, this will provide more immediate, real benefits than any of the above methods.