Next Time: Three Things to Do Differently

By Jon Greenberg
Executive Editor
New Hampshire Public Radio

I admit it. I had big expectations for our most recent interactive news exercise, the "Budget Builder"  at NHPR.org. In fact, at the risk of personal embarrassment, I'll tell you I thought 5,000 people would use it.After all, nearly 30,000 people had participated in our New Hampshire Tax Calculator several years ago.In the end, only 1,800 people built a state budget, and only about a hundred of them shared their budget plans with the public. To be sure, there 's a lot in this project I will copy in the future. There are also several things I'll do differently next time. Let me share what I think went right and what I think dragged down our numbers.

A thumbnail sketch: The state was facing a $100 million to $200 million deficit and we knew the budget debate would be in the news. We wanted to give people a truly clean slate – let them craft a simplified budget from scratch then see how their priorities stacked up to the real thing.
The Budget Builder gave users 11 state spending categories and let them allocate the New Hampshire budget according to how they thought the money ought to be spent. We used a Flash interface so that, as users entered percentages, the slices of a pie chart would fill in.

On the plus side, the interactive pie-chart portion worked great and many people raved about how cool it was. The site was rich in contextual information about the state budget. Users could also send their ideas to elected officials and each other, although only about five people used that last feature. The site let people add their opinions to a survey and see how others were thinking. The navigation was simple and effective. The graphics were attractive.

From the start, I specified to the nth degree the sequence of what the user should see on the screen. The graphic designer and the programmer found these specifications helpful when it came time to make the final tweaks to the project. Three weeks and $4,500 after we began, we were live on the web.

The project was so topical I was sure at least 5,000 people would try it out, especially when I heard that people were passing the link along to their friends.

So what went wrong?

First of all, I don 't think that many people really felt confident about digging into the budget. I had set up a little quiz in the program. People could first create their dream budget. Then they could guess at how the state actually spent its money. Finally, they could see all three budgets – their dream, their best guess and the real thing.

Feedback suggested that many people found the quiz overwhelming and felt they weren't smart enough to play the game. We told them there were no wrong answers; the only thing that mattered was their values. Nevertheless, many were so unsure of what percentages to put in, they bailed out soon after they saw the blank pie charts awaiting their decisions.

I think I asked for too much of the users' time. The typical web experience requires little time: a question here, a little browsing there. In contrast, the Budget Builder asked the user a set of 11 questions – not once but twice. Minnesota Public Radio came out with a Budget Balancer soon after we did and followed a model from the Sacramento Bee that posed a single multiple-choice budget question on each page. That might raise some journalistic issues (who decides the choices, what if the choices in the legislature change), but I think that format works better on the web.

The Budget Builder model reminds me of the historic information posters the D.C. government has put along the sidewalks. They are big and probably well-written, but I wouldn 't know. When I'm walking, I'm not in the frame of mind to stop and read them. Web users don 't take more than a couple of minutes – at most – to complete an exercise.

The lack of quick feedback for users was the second design flaw. I experimented with using the Budget Builder as an opinion poll, complete with user demographics – anonymous, of course. But this function came into the project after construction began, and we missed a key step.
We told users they would be able to see how their dream budgets compared with all the other submitted budgets. But that result should have popped up at them as soon as they submitted their version. As it was, they could see the other budgets, but it took an extra click.

Finally, we should have had a way to plug in immediate displays of the governor's, the senate's and the house's versions of the budget. This would have anchored the exercise to real policy decisions.

There were other issues. The war in Iraq took everyone's attention just as we were ramping up the Budget Builder. My radio station spent a week promoting the new site but it needed much more; I found many devoted listeners had no idea this interactive tool existed.

I think superior design would have overcome some of these problems. Fewer questions per page. Immediate display of results. Obvious connections to the positions of decision makers.

Three good points to bear in mind for the next interactive project.

Greenberg 's work in New Media won the station a national Edward R. Murrow Award for 2003.