Measuring Success in Web Projects
A Ship Without a Rudder
When I was a younger developer, I struggled often with the idea that I wasn't doing meaningful work. Yes, I was building good websites. Yes, they meant something to somebody (at least to the client). But is that brochureware site really going to make the world a better place? Does this university sub-department really need its own website? (I don't think I looked at a single individual department site when I was in school.) When you work on a project that doesn't have a clear and obvious goal, it's easy to fall into this line of thinking. And that line of thinking is very demotivating to a developer - you start to feel like you're spending your days doing busywork. And that's not even touching the fact that web projects are expensive. That university sub-department is probably spending tens of thousands of dollars to make a site that the students may never visit.
The solution is to figure out and clearly define the goals of a site at the outset of the project. The main purpose of that university department site is to serve information. Potential students can learn about the department, and current staff and students keep up to date on what's going on. There are secondary purposes too, like directing potential students to the main school website, getting them to submit a contact form, or getting people to sign up for a newsletter. Just knowing that much gives you something to work towards. If your work doesn't contribute to one of those goals, then things are getting off track.
Getting on Track
Once the list of goals for a site is known, it should be communicated to the whole team. The client knows why they need a website, but the development team may not. Keeping your developers informed of a project’s goals will keep them motivated and on track.
Some projects have obvious goals. An e-commerce site, whose sole purpose is to to sell products, is probably the easiest example. This type of project is deemed a success when the profit from the store surpasses the cost of development. Sites that market an upcoming product may have obvious goals too, even if their value is more difficult to calculate. In the case of a promo site for an upcoming service, the business may plan to collect as many email addresses as possible,assigning an estimated value to each address collected. Then there are sites like libraries and club forums, which don't make any money, but have value in other ways, i.e. providing information and fostering community interaction. In those cases, the value is in making it as easy as possible for visitors to find what they are looking for. If visitors to the library site are getting lost on the way to finding content, the project is in trouble.
In any case, it's important for a project to define success for when the project goes live.
The next step, after knowing the goals, is to refine them into specific user actions like viewing a specific page or clicking a specific button. These "goalpost" actions are defined as something that users can do so we know when they have reached one of the site's goals. This makes each goal both specific and measurable. If the goal is newsletter signups, the goalpost might be "user submits the newsletter form and reaches the signup thank you page." Now every time a visitor reaches the signup thank you page, we can mark a tally because the goal is specific. Looking at the percentage of visitors reaching each goal gives us what's called a conversion rate. The higher the conversion rates, the better your website is doing what it was made to do.
Learning to Measure Progress
Setting up goalposts is what separates directionless projects from meaningful ones. Whether you, as a developer, fully stand behind the product or not, you're helping someone bring their idea into reality. And more importantly, by measuring the value of your contribution, clients can see the return on their investment.
Websites actually have it relatively easy when it comes to tracking goals like this. The most common bit of tracking info in web is just the path a user is on. If a user ends up on a path ending in /checkout/complete, chances are they just finished placing an e-commerce order. Your analytics package will note that /checkout/complete got a pageview, and now you have a tally that one more visitor completed a purchase. Of course, you can also log in to your commerce site to see how many orders are placed, but analytics dashboards are extra convenient and give you endless options for number crunching and filtering data.
For more complicated goals, any decent analytics package will let you submit custom data and track events manually. For example, if you're only tracking checkouts with /checkout/complete, you don't know how much money each individual order was worth. Using this simple method, someone buying a $10 t-shirt looks exactly the same as a person buying a $20,000 giraffe. (Why you have that in your store is beyond the scope of this article.) Custom analytics events let you pass the real value of the order, which might make a big difference depending on what you're measuring. Maybe people entering your site on the homepage mostly buy t-shirts, and people who click through from your Giraffe for Sale banner ad mostly buy giraffes. These distinctions are important for tracking the success of ad runs.
So you know your goals, you've decided on goalposts, and you’ve configured them in an analytics package. Unfortunately, users still aren't reaching the end point. What do you do now? Goal funnels can help identify pain points for your users. A funnel is a list of steps that a user takes on the way to a goalpost (marked by specific and measurable mini-goalposts, of course). Using commerce as an example, step 1 is viewing a product, step 2 is adding to cart, step 3 is starting checkout, step 4 is billing and shipping info, and step 5 is the goal - a finished checkout. Each step has a page or action associated, so we can track how far into the process users get. By watching the percentage of users hitting each step, you can see what step is scaring users off. A huge dropoff on step X tells you that something about that page is bad, wrong, or just plain broken. Revisit that step, develop a more user-friendly way to do it, and then after you push it live, measure the new conversion rate. Since you already have a funnel in place, you don't have to guess whether new development is actually better or not. You'll have percentages to back it up. (Prettier isn't always better. Sometimes users prefer unexpected things - Craigslist has barely updated its style in the last 20 years, and yet it remains popular.)
To recap, by defining and tracking the goals of a website developers stay motivated, clients focus on real sets of goals, and managers can look at the numbers and see real return on investment. So at the outset of any web project, lay out the goals. Once a general list of goals is known, refine them into specific user actions that can be tracked and measured. Analytics software can tell you the value of the completed development. If users just aren't hitting your goals, use funnels to look for pain points to improve on. All of this together will give your projects a more defined purpose with less guessing and more transparency.