Ken Armstrong and T. Christian Miller found themselves in the odd position of moving from competitors to collaborators, over the course of a phone call or two and a few emails. Miller says calling the lawyer of a potential source and finding out that another reporter had already been talking to her “was one of those moments both awful and fortuitous.” After his stomach dropped, he says he quickly decided that rather than try to scoop the competition, there was a chance to collaborate and make a stronger story. On a practical level, that was probably wise: as Miller would learn, it had taken Armstrong six months and scores of calls and emails with the attorney for “Marie,” the victim in the story, before she agreed to talk with him.

It helped that the two investigative reporters work for philanthropic journalism organizations, Armstrong The Marshall Project, and Miller Pro Publica. Both are collaborative organizations to begin with. It also helped that their editors, Bill Keller, editor-in-chief of the Marshall Project and Pro Publica Senior Editor Joe Sexton, knew each other well from working together at The New York Times. Armstrong and Miller also knew each other by reputation and from casual interactions at things like conferences for Investigative Reporters and Editors. Finally, it was a bit of serendipity that the two had been working what turned out to be opposite ends of the same story.

In the end, they produced “An Unbelievable Story of Rape,” which was recently named a finalist for a 2016 National Magazine Award, in the feature writing category. There will also be a segment for This American Life, reported in partnership first with The Marshall Project, and later adding Pro Publica.

Before joining the Marshall Project in 2014, Armstrong, 53, worked at the Chicago Tribune and the Seattle Times, where he shared in two staff Pulitzer Prizes for Breaking News, and won the 2012 Pulitzer Prize for Investigative Reporting for a three-part series he wrote with Michael Berens. He’s a 2001 Nieman Fellow. Miller, 45, worked at the San Francisco Chronicle, the St. Petersburg Times and the Los Angeles Times before joining Pro Publica in 2008. He’s won numerous awards, including a Livingston Award and a George Polk Award for Radio Reporting. He was a 2011 John S. Knight Journalism Fellow at Stanford University.

The two spoke with me via phone from their offices in Seattle and Berkeley, Calif. This interview gives away the ending of their story, so you might want to read the story first. Their organizations have also published pieces on how the two swallowed their scoop instincts to work together and how their organizations worked to collaborate. Armstrong and Miller put together a detailed discussion of their sourcing and documentation, and a primer on the best reporting on rape. We focused on their collaboration. This interview was edited and condensed.

How long did the writing actually take [for the 15,000 word first draft]?

Armstrong T and I are both pretty fast writers. It’s hard for me to go back and reconstruct in terms of days or weeks how long it took. You’re writing one chapter, then you’re doing some reporting. You write another chapter and then do some reporting. But if I had to estimate, I think for the Washington chapters all told it was probably a couple of weeks for a first draft. Does that sound right, T?

Miller I started writing a little bit earlier than Ken because he was dealing with some other stories he had to do. From the moment I touched the first keyboard to turning in our first draft was six weeks to two months. But that wasn’t straight through writing. I would write for a while, then you started writing for a while. Then I went back out reporting, then you went back out reporting.

How did you come up with the narrative structure?

Armstrong I had a structure in mind that I really liked, I thought it was quite good. It was a braid, you went back and forth between Washington and Colorado. We talked about it [at a meeting in New York, about two weeks after the two organizations decided to work together]. The next day, before I got on the plane to come back to Seattle, T called and said that Joe had an alternative idea in mind for a structure, and wanted to know if we could meet again. T and Joe came over (to the Marshall Project offices) and Joe described his idea for the story structure. And immediately upon hearing it, it was clear that however good I thought my structure was, it wasn’t nearly as good as Joe’s structure.

Miller I remember we had lunch at that crappy burger place, it was supposed to be a hamburger joint, but it was in a five-star hotel, so it was a faux hamburger joint. A $22 hamburger! We had a braiding but it was two or three or four sections. It was the beginning and ending where Joe had different ideas, but Joe’s was much more kinetic. It had a lot of energy in it, moving back and forth rapidly [ultimately, it would shift between Washington and Colorado nine times]. The obvious way to start was to do a chronology and start with the rape of Marie. That was the most powerful moment. Joe said, you didn’t want to start with that, you wanted to build towards that. And that was a powerful insight

Armstrong You heard it and you knew it would work. You could have started the story in half dozen ways, you could have ended it in half a dozen ways, but his structure preserved suspense and it worked toward an emotionally wrenching end in a way that was really, really powerful. It’s always helpful when as a reporter the editor comes up with the structure. Because when the reporter comes up with the structure, there will almost inevitably be, um, negotiations at the end (laughs). In this instance, when you’re writing towards an editor’s structure it tends to make things more efficient. And it worked. It worked beautifully.

Did you work with one editor primarily?

Armstrong There was a lead editor from each news shop. Joe Sexton was the lead editor for Pro Publica and Bill Keller was the lead editor at The Marshall Project. But numerous other editors contributed at different points in the process.

Miller Most reporters’ blood curdles at the idea of multiple editors, but each editor really brought something to the story. It wasn’t painful to me to go through three or four different editors. It really did lift the story at each stage. I’m not saying this to be nice. I think Ken would agree that at each stage, the editing improved the story. I don’t always feel that way.

Armstrong I have to agree with T, layer upon layer of editing doesn’t always benefit a story. But it did in this instance. And I remember late in the process, after it had been edited multiple times, Kirsten Danis, the managing editor at The Marshall Project, went in and did this precision edit for the first two sections of the story. And it was so good that both T and I individually asked her to keep going and edit the entire piece. Then we could incorporate what we felt was really needed. That’s pretty rare when you have a story that feels like it’s been done, that’s been put to bed, and then another editor comes in, and both reporters say, “Yes. Please keep editing us. We would like more.”

Did you merge your reporting, or did you each focus on the story you’d been working on, then come together to write the piece?

Armstrong It was almost organic. I had done the bulk of the reporting on Washington and he had done the bulk of the reporting on Colorado. It made sense for us to take the lead on those different parts of the story. We just basically divvied it up. I would report out what happened here, he would report out what happened in Colorado. In doing the weave, we had to know what the other was writing. Not to emulate or match each other’s voice, but to avoid continuity errors and to make sure the story as a whole made sense. So we shared chapter drafts as we went along. We also traded ideas on what to reveal where, in order to help the narrative build.

Miller Part of the magic is that we were writing blind mostly. We were talking all the time, but Ken would go into writing and I would go into writing and then we’d sort of combine. We had a structure we were writing to, but how we got there was unto ourselves. When we ended up mixing together, it was startling to me how nicely the two voices fit together, the two stories flowed together. I wouldn’t have necessarily put my money on that working. But it did.

Would you two merge your copy and then send it off, or were they editing your sections knowing that there was this braided structure? Who was in charge of the braiding?

Miller The early drafts we [Miller and Armstrong] were swapping back and forth. We sent jointly to the editors a complete story, a whole story, woven together. From that point forward it was edited as a whole. Joe and Bill and Kirsten or Robin Fields [managing editor of Pro Publica] would look at the whole and return the whole to us. In Ken’s section of edits, he would deal with those edits, and in my section of the edits, I would deal with those questions.

Armstrong There were revisions as it went along. One of the things Bill felt strongly about was he wanted to make sure we were economical at the story’s end. He wanted to make sure that we were ending on a high note, and it didn’t feel like we were dragging. Our original draft I think had 15,000 words. Bill was like, “No, this is too much at the story’s end.” So that very short compressed epilogue that we wound up with was really a product of Bill and other editors saying, “We’ve got to take this 3,000 words at the end of the story and compress that. We need to let the reader go now.” And that was kind of the dynamic throughout.  T and I would also tinker with the different chapters in terms of where we would end it. Sometimes as we were writing, we’re sharing a draft back and forth, we would find a better exit point or a better entryway into a chapter.

Miller About the reveal [of the actual rape], this is sort of comical, too, as we look back at it. I don’t know how much sleep Ken and I lost over whether or not we were going to give away the ending [too early], and when we would give away the ending. By the end I was convinced that any reader would know by the first section at the top that Marie had been the victim of [an actual rape]. I thought, ‘We can’t pull this reveal off.’ It was interesting to me from the perspective of a writer who’s writing a story and thinking, ‘This isn’t going to work.’ And actually it does work, for some people anyway.

Armstrong We really tried not to give away the story before people had an opportunity to read it. We had buy-in from everyone in terms of the headline, the cutline, the art, even social media. T and I were both surprised at how well it worked. Most readers, it seems, weren’t sure how the story was going to end before they got to the final chapter. And I’m with T, we both thought that people would see where we were going 100 yards before we got there.

What was the biggest snarl that came up while you were working together on this story?

Armstrong Somebody at Pro Publica came up with the idea of T and I writing separately, to make sure we weren’t trying to merge our writing voices. I could write the Washington chapters and T could write the Colorado chapters. That prevented a lot of the obstacles that you would normally see in a collaboration like this. T and I have distinct writing voices, but the differences aren’t jarring. In this case I think each of our styles suited the subject material. T is a more descriptive writer than I am. His writing has almost a cinematic feel to it, and that was really so well suited to what he was writing about, which is this sprawling tense chase, the search for a serial rapist. My writing tends to be a little more spare, and what I was writing about was a study in doubt. How doubt sets in, how it spreads. I think a more muted voice served that subject well. I think one of the big debates we had was how to write the final scene. T, you want to talk about that?

Miller The biggest overall concern with the story was ending it in a way that didn’t feel gratuitous, that didn’t feel like we were being disrespectful or salacious in recounting the most awful experience that one could imagine occurring to a young woman at 18, of being raped for hours by this stranger. We went back and forth, and decided things like, “What do we want to say here? What details are too much or too little?” That took quite a while to figure out, how we wanted to write that section, what exactly the tone was going to be. The writing care came to its peak in writing that last section.

What’s the biggest challenge in reporting on and writing about sexual assault?

Miller I mean, everything. Ken, why don’t you tell him the Jim Neff story?

Armstrong My former editor at the Seattle Times, Jim Neff, writes non-fiction books. One book he wrote years ago dealt with a rape case. I remember Jim telling me on more than one occasion that there is such fatigue with reading about rape. There’s something about that particular crime that is just so personal and invasive, it’s just something that people don’t want to read about, particularly at length. He was saying he would never again do a book about rape, that it was almost impossible to write about that subject at book length. T and I, what we were writing was not book length, but it was close to being novella length. We were writing about a serial rapist, we were writing about a half-dozen sexual assaults or attempted sexual assaults. And as T said before, we’re trying hard not to be gratuitous. We don’t want to be graphic unless it serves a purpose. That was one of the great challenges.

The other [challenge was], T and I are both men, the two lead editors were both men. We wanted to make sure we had input from women readers as well, who might have a perspective that we didn’t have, and might help us if we were going down a road that really, we didn’t need to go down.

I think everything about it is difficult. We talked at length about how we were going to write about specific themes and what the tone would be. And we tried to share with as many people as possible beforehand, trying to get feedback and see how people reacted to different sections.

It sounds like this is a difficult topic even in the realm of investigative reporting?

Miller Aside from child sexual assault, which is perhaps the most horrible [thing to report], yes. Everything is difficult about the reporting. It took Ken six months of making his case and trying to bring forward the best case for writing the story until Marie was convinced he could do it in a way that was sensitive and fair and would be of a help to other people. The stigma surrounding it makes it so that women don’t want to talk about it particularly, nor do men who’ve suffered sexual assault. If you’re going to have a story that makes the police look bad, obviously the police aren’t particularly happy to talk to you about this. It’s one of the few topics where really nobody wants to be in print about what happens. I’ve done war crime stories, and stories involving people in the worst moments of their lives, and nothing really compares quite to the degree of difficulty in doing a sexual assault story justice.

Armstrong What really served the story was the depth of reporting. We interviewed Marie, who had never before spoken about this. The lead detective in Washington who botched the investigation. We interviewed the serial rapist. We interviewed the two detectives in Colorado who helped solve the case. In addition, we had thousands of pages of records documenting how the investigation went right in Colorado and how it went wrong in Washington. With very few exceptions, the people we most wanted to talk to we were able to talk to. I think that the writing of the story succeeded can be traced to the depth of reporting.

Miller It is a really rare story when you’ve got the perpetrator talking to you, the police are talking to you, the victim is talking to you and the police who didn’t do as good a job also decide to talk to you. Each of those was a reporting coup in its own way. For all to materialize, as Ken said, it’s a testament to the depth of reporting and what it can do.

I was impressed that the detective in Washington chose to speak with you. Was that as hard as getting Marie to talk to you?

Armstrong That happened very late in the process. For months we were getting a “no comment” from the Lynwood police department. The interview with the lead detective took place the day before we published. So we did the interview with the Lynwood police and then I came home and incorporated what we had learned from Sgt. Mason in the story the afternoon before we published.

Miller I was involved in getting the story in the fricking paper. I felt like I was at the helm of the starship and Ken was on a mission to the planet. Because it wasn’t just Sgt. Mason, we also got [two] upper echelon guys in the police department, and you got another one, right, one of Marie’s friends? Ken spent the whole weekend interviewing. There were copy edits and the photos had to placed, and Ken and I would have these conversations where we’d briefly consult and he’d come in and fill in some stuff in the story, and he’d go back out for another interview. It was good that we had the relationship we did, because we both had a lot of trust in each other as reporters and as writers. And we were able to pull that off.

Armstrong I came back and was writing these inserts, and all four of us are on the phone together. T’s in California, I’m in Seattle, Bill’s in his office in New York, Joe’s in his office in New York, and we’re all on a conference call and in a Google doc, looking at the changes together. In just a matter of a few months, I think we came to a point where we can work well together. That could have been a real challenge, to incorporate changes that late in the game. But it really wasn’t. It actually came together pretty quickly. And it worked smoothly.

How did that last minute interview with the detective change the storytelling? What was it like to have that happen at the end?

Armstrong I give the Lynwood [police department] credit for owning their mistake. To my mind, people are more forgiving if you are upfront about what you did wrong, and that you feel just awful about it and are determined to prevent it from happening again. In those interviews, they took the opportunity to make all those points in a way that was really genuine. They are doing some things differently now. And when you sat in the room across from Sgt. Mason and you heard him talk about the mistakes he’d made, it was really hard for him. It was very emotional. And it was clear that this has had a profound personal impact on him.

It’s something rare. A week or two beforehand, we finally got these internal reports. One was a peer review report, which looked at the various things the Lynnwood police department did wrong. Then we got an internal report [from Lynnwood]. Those reports really helped elevate the story. Getting that additional documentation really helped us detail what went wrong.

Miller I want a plug here for public records, and for using public records to sort of cross examine each other. This story would have been probably impossible if it had taken place in states that don’t have good public records laws. Both Colorado and Washington were responsive to public records requests. That allowed us to have a lot of the details, a lot of remarks that were made, and the scenarios. The reason I think the Lynwood police [talked], one is we had at the very end found this internal report that they hadn’t provided to us. Ken was able to use that to say, “Hey, what are these reports about?” I had been tracking down another Lynwood detective in California, and I had been increasingly ratcheting up the pressure for him to talk. He wouldn’t talk, but he was forwarding emails to the Lynnwood police and saying, “Hey, this guy is honing in on me.” I think that combination of that pressure finally shook them loose and convinced them it was worth talking to us. You think that’s right, Ken?

Armstrong Once we had that internal peer review in hand, I think that may have helped play a part. Also, this was a situation where a particular commander there found out what we were trying to do, and I think he was given to the idea that we [the Lynnwood police] need to own our mistake, and that the best thing we can do is talk about what we did and what we now do differently. I think he really helped swing their thinking on this.

Do you see yourselves collaborating again as reporters?

Armstrong We both work in places where the impulse is to collaborate and not compete. I could easily see our paths crossing again. If so, the fact that this worked out so well I think will make it easier going forward. But collaborations don’t always work out well. We were very fortunate in this instance. Part of it was that T and I already knew each other, Bill and Joe know each other well. There was a familiarity here that did not breed contempt. Instead, I think it helped breed a certain confidence that we could all work with each other, and the story would benefit from it.

Miller A collaboration that ends positively usually leads to more collaboration, in my experience. Ones that don’t end positively don’t lead to more collaboration. This was a super achievement on both organizations’ parts. That it came off as well as it did I think bodes well for the future. If there’s any bigger lessons to be learned, I’m not exactly sure what they are. A lot of this was sui generis. It happened that we had the same story, it happened that we reported the stories at the opposite end, it happened that we knew each other, our editors knew each other, and we’ve all worked together. I would like to believe this was a highly reproducible pattern. But I’m not sure. Time is going to tell on that one. It certainly was a different partnership in terms of other partnerships I’ve worked with. Usually it starts from the beginning and you agree on how to go forward. This was a shotgun marriage that ended up working out really well.