The fact that we were working uphill into a headwind actually benefited the creative, as we had to make everything rock solid. Joe Sabia and I sat down and we asked ourselves ‘how does this make sense? How much does the video cost? How much do all these overheads cost? What do we have to be doing to make this make sense financially?’

At the time there were Condé Nast branded videos getting around 20,00 views, so we thought - okay, if you have 20,000 view videos going out and we turn them into two million view videos, you create a hundred times more efficiency for each video.

But the thing is, we didn't do it in one step. We 10X'd it and then we 10Xed it again - once we’d 10Xed it in watch time, then we could start selling midroll ads and then the sales team started to take notice.

The best thing about Youtube and its big mad friction is just one simple thing - do we know what people like? The friction is not some deep and arcane thing about understanding the platform and ‘what the algorithm might want’. All that to me is just boring, optimizing - very dull.

The big friction is, can you work out what people like? With every celebrity that would come through, we would be working out what do people most want to know about them? And that's what you design around. You can't have a massively successful show if you're not directly hitting the core appeal of the people that are coming in.

I used to use the phrase ‘structured play’, which is just, ‘let's work through this game’. I also had a phrase ‘skip in public’, which meant you should let the contributors say ‘not answering that one!’ on formats like AutoComplete Interview. I like it when the celebrity just tears through the questions, when they go, "Not telling you. Nope, that doesn't even make sense.” And then stop on ones they want to spend time on.

I'd often say the difference between a TV show and an internet video is a TV show is highly structured in advance and presented out, whereas in an internet video, you make something happen and you capture that, you capture the fact it occurred. We’d create an experience or an experiment in the format, and then record the results.

Our biggest bet early on with the non celebrity expert videos was just to go deeper than anyone's gone before. I think a really great example of this is a format that Joe directed where an accent expert breaks down accents from movies and he just goes to an insane level of depth. He's talking about the non-rotive R and vocal fricatives and taking it to a place of using something very familiar to take you to something very unfamiliar. The more familiar the initial thing is and the more unfamiliar the place you take them, the better.

I believe that turning bad friction into good friction is a real art. Good friction implies agency - something you can address within your team, but the bad friction is outside. So to turn bad friction into good friction you have to bring it in closer to your agency.

One of the biggest sources of bad friction is when people are not necessarily honest about goals. They have their own goals, which may be counter to the conversation you’re having as a team. I do a lot of hiring for brands making content now, and I mention it in almost every interview - what are your goals, and are our goals and the brand’s goals aligned? Before you set out on a journey, make sure you're going to the same place!

For example, a big friction is paying lip service to wanting to make a valuable video format, but valuing your first class flight to the shoot more than the business impact. These are people who aren't trying to make their video into a grand success, they are people who would happily make a video no one watched if it got them to their personal goal.

I think the positive side to all of that is finding points of convergence where goals can meet, rather than saying, "Well, these goals are different and therefore we'll constantly be pulling at each other." It's like, "I think we can find a way for these goals to meet." I think the editorial desires of our brands were always built into the programming approaches, but it was really difficult for all the stakeholders to see it.

Finding that convergence, turning bad friction into good friction, involves getting creative. Now I always do this chat with clients, which I think of as the goals, limitations and values conversation, and I’m just really clear about asking that question. And if you can't clearly say your goals, that's what we need to be working on, that’s what we need to apply our creativity to first.

“Somebody once said to me, "Yeah, well, the problem with this is it's a bandaid." And I was like, "It's bandaids all the way down!" It's this great big ball of bandaids where everything was made as a temporary thing, and then it became a standard thing, and then it became a central thing.

I believe in transparency as the route to accountability and accountability is the route to efficiency. And as long as you can have candor and kindness you can then just root the thing out and go, "I know it's messy, but it's time to open the box of cables that you've got and sort through them." Another metaphor I used was setting up floodlights, just going around and setting up lights because essentially where it's dark, things will get put away out of sight, things will get hidden. You have to have that kind of militant attitude to get it all out in the open.

When you're able to talk honestly about these things, then lots of things can get unhidden. Sometimes you have to reckon with things like being able to have a chat with one of your creative team, and say if something isn’t good enough, rather than saying its ok. Every now and again the floodlight is telling them “you need to get better at this, but let's work together, identifying where are you weakest and what can we do to improve?”

I think that’s what I love about working with YouTube -  it's unforgiving and it's honest. You upload the video and it either works or it doesn't. And that's why I really don't like people just going, “oh, it wasn't my fault. It was the algorithm” or “it wasn't my fault, something fell apart in production.” I think that's where you can set up floodlights in a creative process and say, “are our ideas any good? Do we need to do this in a different way?”