Everything I know about structuring a feature

From my advice to the CAJ mentorship program

1. A feature tells you one thing, in detail.

In 1994 (I know! Long time ago) I wrote my first magazine article, for a defunct magazine called Saturday Night. I was starting to write about politics for the Montreal Gazette, so I wrote about Jacques Parizeau, the separatist Parti Québécois leader, who at that point was about to be elected Quebec’s premier.

I set aside a month to write the article. The magazine’s editor made me spend half that time refining my pitch — my argument about what the article would say. This seemed like a terrible waste of limited time. I finally realized he was teaching me how to write a long article.

In successive pitches, I kept adding new elements to my description of the article’s many wonders. He kept saying, “But what’s it going to say about him?”

In desperation, I bought a bunch of magazines and read a bunch of features. I’d been reading magazines for, at that point, nearly a decade, but I hadn’t really analyzed what I’d been reading from the viewpoint of structure. The main thing I noticed was that a long story has one big point to make about its subject, just as a short story does. The long story just has more detail, or more arguments, or more context. But still one point.

There are countless exceptions. I never did figure out what point Susan Orlean was making in “The American Man at Age 10,” except maybe the child is the father to the man. I think she was trying to say a lot of things, or nothing in particular. But she’s great and that story will last forever, so whatever. But way more often than not, there’s a single argument. It was negligence that caused the disaster. The candidate isn’t ready. The candidate is ready, despite appearances. The elaborate plan is doomed. Our heroes succeeded despite their doomed plan. The story of my poor aunt helped me understand life. Or the story of my poor aunt kept me from seeing what really matters.

On Parizeau, who was a doughy and cheerfully pompous economist who sounded British even when speaking French, my argument was that he was by far the sovereignty movement’s most formidable figure. Already a cheeky claim, because dashing Lucien Bouchard was already getting more ink. Later events would seem to prove me wrong, but I stand by my claim. Whether I was right matters less than that I had a claim, because having a central argument helps you make everything else coherent. 

Everything you’ve gathered for your feature – interviews, archival research, structural thinking, observed colour – must be compared against the point you’re trying to make, and discarded if found unhelpful.

You don’t need to know what your point will be when you start working. In fact it’s usually best to let your work strongly influence your choice of a point. (Or a thesis if you want a fancier word.) You should definitely decide what your point is before you start to write, because writing is a process of deciding how to illustrate and defend your central claim. And you’ll probably still be reporting when you realize what argument or claim you’re going to be making. But don’t worry if you’ve done a lot of reporting and you still don’t know what your article will be about. You’re exploring.

2. Picking a structure is the second-last thing you’ll do. The last thing is to write the feature.

What comes before you choose your structure? All of your reporting. Or as much of your reporting as possible before deadlines loom. That means your reporting should come first — before you’ve spent much time deciding how you’re going to tell your story. Structure is a way of deploying the information you’ve gathered. So your choices about structure will largely depend on what you’ve gathered. Picking a structure, an argument, a claim before you’ve done your reporting amounts to dismissing, from the outset, the possibility that your research might surprise you. Which is a good way of ensuring nothing you write will surprise your reader.

Research can take many forms. Interviews with the subject, their friends, their antagonists. An institutional paper trail, like court rulings or relevant statistics. Reading for historical context — When has this sort of thing happened before? — and comparative context — Where else does it happen? (I can just about guarantee that whatever you’re writing about has happened before or somewhere else. Finding that context will set you apart from all your colleagues who will claim everything is brand new, simply because it’s new to them.) Even analogies to literature or to academic research can be pertinent. 

Once you’ve got your research, gather it in one place, re-read it, and take notes on the themes that emerge. He had a rough childhood or Officials in his position often crack or His favourite teacher once predicted this would happen. Consider different ways to deploy these elements, different orders, different emphases — always in support of your central claim, or thesis. 

The order in which you deploy your story elements will decide your story’s structure. Probably most of your elements will take place in the story’s “now” but there are some parts that take place in the story’s prehistory — the subject’s childhood, say, or the 1840s. You can decide where those background bits best fit, in relation to the more contemporary stuff. Maybe you’ll need a flashback, so most of the story takes place “now” but at some point you segue: “It wasn’t always like this….” Where should that happen? Now’s the time to ponder these questions.

Making these structural decisions before you write helps ensure that an editor won’t need to perform major surgery on your story to fix wonky pacing. Making these decisions after you report ensures that your decisions are informed by a rich understanding of what you’re trying to say. 

3. Any structure is allowed. But any structure has to serve your highest goal: clarity.

Journalism isn’t fair. Your reader owes you nothing. At the first sign of trouble, they’ll turn the page, close the tab, move on to the next distraction. The obligation runs all one way: you have to make your writing as easy to read as possible.

That doesn’t mean you need to make it simplistic. I’m proud that I cover complex topics – budget math, theoretical physics, federalism – but I learned early that when I’m writing about something tough, I need to take the time to walk readers through the material. They’ll stick with you through that. But if you suddenly vault them into a flashback without explaining clearly that you’ve changed time frames between paragraphs, they’ll leave. 

So don’t indulge complex structures for kicks. And if you do depart from, say, strict chronology, signpost the hell out of it. “It wasn’t always so easy. When Jones was 23 she was a graduate student at the University of X. She was immersed in the study of X Y Z.” You should hear the gears shift as you read. I often number different sections of an argument or chronology. “First, blah blah blah. Second, bleep blorp. Third…” Nobody’s ever complained that my writing is mechanical or whatever. At least not because I number things. 

4. Be comfortable making big cuts and leaving big parts out.

Structure is about choice. Dumping your notebook into a Word file isn’t a choice, or at least not a good one. A great writer at The Gazette in Montreal once told me you should use about 25% of what you gather for a long story. That’s not a rule, but you should be prepared for the feeling of putting real work into some aspect of a story, then leaving it out because it doesn’t pull its weight.

My second Harper book was going to be, in large part, a book about Harper as a wartime prime minister. Afghanistan turned bad as soon as he showed up. (Basically coincidence.) There was a whole ethical arc. I ended up writing almost none of that. I just had some other stuff to say, and the book couldn’t be 800 pages.  Incidentally nobody’s ever told me, “Great book, except it was skimpy on Afghanistan.” If your story is about something, most readers won’t notice what it’s not about.