Thank you for taking the time today, Russell. I really appreciate it. To start off, could you tell me your name and what you do?
My name is Russell Samora, and I am a data journalist at The Pudding.
Can you provide more information about The Pudding?
The Pudding is a digital publication that creates data-driven visual essays and stories. While some are visual explorations, most are data-centric. It serves as a glorified staff blog, with team members sharing stories they're passionate about. We also collaborate with freelancers, offering a platform for others to tell their visual stories.
How did it all begin?
The Pudding started in January 2017. Matt Daniels, Ilia Blinderman, and I were working together under the name Polygraph, which is now our creative agency. We wanted to distinguish between agency work and passion projects, so we established The Pudding as an independent editorial publication while keeping Polygraph as the agency. The same team works on both, but under different names. Our revenue comes from the Polygraph side, and The Pudding serves as a fun outlet.
How does your team approach working on Pudding projects?
Our team dedicates 50% of our time to working on Pudding projects, similar to Google's 20% time concept. We strive to sustain The Pudding so we can continue doing these projects. Our process might vary among team members since we encourage everyone to work in the way that suits them best. We don't have strict deadlines or time allotments for stories.
So, can you walk me through your process on a story from start to finish at The Pudding?
Sure. I'll try to generalize it a bit, but we emphasize individual work processes. We don't have deadlines or specific time allotments for stories, allowing authors to work at their own pace. We trust the staff and authors to tell the story in the way that makes the most sense to them. Our process leans into author autonomy and collective accountability. We have checklists that help authors consider various aspects of their story, like getting regular feedback and conducting accessibility audits.
In terms of the process from start to finish, it's loosely defined and is designed to work with the individual's preferred workflow. We have weekly meetings called "story time" where people pitch ideas and discuss them.
How do you provide feedback to ensure a story's success?
It depends on the phase the story is in. I believe giving more feedback early on is crucial, as it helps narrow down the countless possibilities. It's important to make people feel comfortable seeking feedback even before their work is polished. Once they make decisions and close other doors, they can focus on the chosen direction. My feedback aims to help people overcome initial hurdles and achieve small wins to move the story forward.
When you get later in the process, let's say I've already started visualizing, and you're now looking at giving me feedback on interaction, animation, and visualization, what's the common thread or the ethos behind your thinking?
We usually let the author dictate what feedback they need and what's on or off the table. Especially in larger organizations, you often get stakeholders who come in late and suggest changes without understanding the project's backstory. So, we use what we call the "Pixar approach": the author specifies what they need feedback on, people provide feedback on those aspects, and then the author decides which feedback to incorporate. This way, the focus remains on the author's vision and specific areas they wish to improve.
Wondering how you would critique your own work before you shared it for feedback.
That's a good question. I'm currently in that position with a story, so I'm trying to think about what I would do. I guess it depends on which facet we're talking about, but I usually focus on getting the core features in place first. I like to ensure that all the essentials are there so that people reviewing it don't feel like they're missing important information. For me, it's about providing enough context for reviewers, even if some design aspects are missing. The priority is ensuring they don't have questions of understanding.
So, how do you measure impact, engagement, or readership on your stories?
We don't really measure these in the traditional sense. Instead, we customize the measure of success for each story based on the author's goal. For instance, if the story is about musicians, a successful story might be when a certain musician tweets it out or shares it in their newsletter. We have specific measures for what a successful story might be. Technically, we do have views, which come from server logs, as we no longer use Google Analytics or any other analytics. So, the focus is more on monitoring what the author is most keen on monitoring, whether it's social media-related or something else. We definitely prioritize this approach over more traditional metrics.
If you don't have a particular measure across the board for what you think will be an impactful visualization of a given story, what is the ethos? What determines whether you think a story is ready to be published?
Two-part answer. First, I feel like stories are ready to be published way sooner than most authors think they are because we get so focused on polishing and adding extra features when in reality, we don't need all those bells and whistles. As authors, we often feel the need to share everything, but consumers of the story just want a small fraction of it. Analytics have shown that most users don't look at all the extra interactions and tooltips; those are for power users.
So, broadly speaking, a story is probably ready to go sooner than we think it is. However, we lean into the authors' preferences. If an author wants their work to be super polished and perfect, that's their prerogative. But if someone wants to publish sooner and maybe iterate later, that's also fine. We have a different philosophy when working with someone else's content, and from my personal viewpoint, it's up to the individual.
In terms of success by traditional metrics, we've found no rhyme or reason whether publishing sooner or later, or spending extra time on a story, makes a difference. We try to remind people that if they're spinning their wheels at the end of a project, it's probably not necessary, and they can just get it out there and it'll be received just as well without adding extra features.
It's very different from what I've heard in a lot of other places, and it's refreshing. I assume a lot of your learnings are tacit knowledge, context-specific, and hard to articulate without a given context. But are there any generalizations of things you've seen over the years that work or don't work?
Definitely some broad ones. For instance, any opportunity to personalize a story inherently makes it more accessible to a user who may or may not have any vested interest in the story. That can involve localization, or making it relatable to the user based on their location or age. Personalization tends to engage people earlier, even if they're not initially into the topic.
Another one is when discussing a topic that isn't already understood by everyone, I'll usually dedicate the first portion of the article to explaining that metric. I'll never start with all the data and trends; instead, I begin with the smallest atom of the data and explain why it is what it is, putting it into context. It's a build-it-up approach, and I find it to be a universal and clearer way to help readers understand the data behind the story.
That's a good point about making content accessible to different levels of knowledge. And one thing that seems consistent across your articles is that they always address a popular, relatable subject but from a completely different angle.
That's true. When we think about doing new stories, we usually consider if it's something people want to discuss and talk about, like sports or music. But then, are we bringing something new to the conversation? Are we introducing a new data point or a different way of looking at the data? We try not to do anything that people wouldn't want to talk about.
One last question, how do you think AI-powered tools will simplify the process of visual storytelling? Have you thought about that or seen anything interesting lately?
Great question. I follow and play with AI tools. We've done a visual story with GPT-3 a couple of years ago. I use GitHub Copilot, which is interesting. For me, AI is another tool in the arsenal, but it has a learning curve. I haven't had any experiences where it's a drop-in solution. With coding, it tries to predict what you're going to type, which saves time, but you still need domain knowledge or review it. I've tried using GPT-3 for story brainstorming, but I haven't found anything good yet. I'm not sold on it being more than a helpful tool in the arsenal.
That's how I see it as well, as a helpful tool. I was wondering how you were using it in your process.
When I feel uncreative, I see what its responses are, but I haven't found anything interesting from a storytelling or creative standpoint.
Agree. But thank you very much.
Of course. No problem.