200 – Behind the Scenes of The IDEMS Podcast

[00:00:07] Santiago: Hi, and welcome to the IDEMS podcast. I am Santiago Borio, an Impact Activation Fellow, and I’m here with Johnny McQuade. I’m not quite sure what your role is, Johnny.

[00:00:20] Johnny: My job title is simply Software Developer, I think.

[00:00:24] Santiago: Okay. I think you’re much more than that and you discussed that in another episode, so let’s not get into the details of that. I think we all do a lot more than what a role implies.

[00:00:36] Johnny: Yeah, if we had descriptive job titles it could get a bit long-winded.

[00:00:41] Santiago: Yes.

So today’s a special episode, it’s our 200th episode, uh, which is quite a milestone for a podcast, 200 episodes. It’s not a trivial number of episodes. So I’m very excited to be here with Johnny to kind of go back to first principles and tell the story of how this came about and what we had to do to get this going.

And of course we should mention Lucie as well, who has been a crucial member of the podcasting team, but unfortunately it was difficult to find the time for the three of us to record together.

[00:01:26] Johnny: That covers the team who are kind of working on the production of the podcast, doesn’t it? You and Lucie and I are the ones who, that’s why we’re here to discuss, the ones who on every episode process them, edit them, review the script and publish them.

[00:01:46] Santiago: Yeah, and I suppose part of the objective is to do a sort of behind the scenes and explain what we do, what each one of us does, and how things evolved through these 200 episodes, because I think if you compare the quality of the episodes now to the first few, the difference is quite noticeable.

[00:02:06] Johnny: Yeah, for sure. Maybe it’s good if we start by talking about the origins of the podcast and why we decided to do one at all.

[00:02:15] Santiago: Yeah, I think this came from David and Danny having a chat after a networking event that David went to. David being of course one of the founding directors of IDEMS, he went to a networking event where someone he got talking to was very impressed with everything IDEMS was doing and said, well, where’s your blog? I want to read more. I want to know more.

And David said, I’m dyslexic, I can’t write very well, which I don’t think is a fair statement from him, he can write very well. But, he was then discussing with Danny and Danny said, well, you’re not too confident with your writing, you’re very good at speaking, why don’t you just record and do a podcast?

And they presented the idea in general to the whole IDEMS team and a few of us said that we were interested in exploring it. I think the two of us, Lucie, were the key people who suggested would be interested. I can’t remember if anyone else.

[00:03:30] Johnny: Yeah, there was people who were definitely interested in playing a part. Lily was involved very early on.

[00:03:36] Santiago: So people were interested in being participants in the podcast.

[00:03:41] Johnny: Yeah.

[00:03:42] Santiago: But the three of us distinguished ourselves from the rest by saying, actually we want to work to make this happen. And you went off to do a bit of research.

[00:03:52] Johnny: Yeah, I think my interest came from the fact that I listen to a lot of podcasts, it’s probably the medium that I engage with the most, even compared to TV or reading or films.

[00:04:09] Santiago: Yeah, same here.

[00:04:11] Johnny: Most of my information, yeah. So I kind of had an interest in how it might be, how podcasts might be made, and I felt like I had an idea of what an end product could be like. And I think part of that was, ’cause in those early days, we were clear that we didn’t want to spend too much time trying to like optimise things like sound quality, we just wanted to get something out that was good and that we absolutely backed, but we didn’t want to kind of spend ages making things overly polished.

[00:04:47] Santiago: Sorry, I must clarify. You said something that’s good. I think the word or the phrasing that David used was something that’s good enough, not necessarily good, but something where we could, in a sensible way, tell some of our stories.

[00:05:03] Johnny: Definitely, yeah, good enough that we’d be proud of it. I think we knew we had a lot of stories that we wanted to tell, and I suppose that was kind of, yeah, seeing as an alternative to something like a blog where there’s loads that IDEMS is doing and there was no documentation of it in any form.

And so having a kind of short pipeline of getting these conversations into a public realm was just really well suited to us. And I think that, yeah, it was quite fun knowing that good enough was the aim, and so we really tried to tighten it up to being as short as possible in terms of the production time for each one.

So if you wanted to, we can be responding to things that happened in the news that week or things like that. And maybe it’s worth saying as well, David was clear from the start that we wanted to do two a week, which to me seemed like a lot initially and potentially a lot of extra work to be managing that, considering that, you know, it didn’t have any tangible goal or, yeah, making anything, getting anything back for IDEMS necessarily.

But it hasn’t been difficult to find two conversations each week that are worthy of being shared.

[00:06:14] Santiago: Well, it hasn’t been difficult. You say that, we have had moments where we had to push people to get out of their comfort zone and actually have those discussions. There’s a couple of people who were very hesitant at first on recording episodes, on being in the podcast. And then once they started they realised, oh, actually this is not that bad, we might as well do a few more episodes later on.

And we started adding voices to the team because initially it was David, of course, Lucie, Lily and myself doing pretty much all of the recordings. And then we included Kate when she joined, George came in, you came in as well with a few episodes. Then we started having external people to IDEMS as well, both from our partner organisations in Kenya and Ghana and from different projects that we’re doing, like the research method support in West Africa, we had quite a few people from there that David interviewed.

We had a couple of completely external people that had topical things related to our work that David interviewed as well. And it evolved to become what it is today where we describe in our conversations our work sometimes, and sometimes, as you said, topical issues that are happening in the news, sometimes we look at thought experiments. I particularly enjoyed two recent episodes I did with David on gamification of education that is completely thought experiments that align with some of our principles and values.

But let’s go back to the beginning again. David said let’s get a podcast done, and he got a team together for working and said, okay, why don’t you do some research, estimate how much time it would take, what it would require, and let’s see if we go ahead with it. And I believe you did that first initial research.

[00:08:35] Johnny: Yeah, I think that’s right. I was looking to the kind of technical side of things of how it would be possible to both edit and publish, and with the kind of thought in mind of making as little extra work for ourselves as possible. The recording was always going to take a certain amount of time, but we wanted to minimise, especially as it was two episodes a week, the additional time needed from the team to get that out. And this was, well, was this 2023? End of, was it two years ago now?

[00:09:11] Santiago: Well, it was a hundred weeks ago.

[00:09:14] Johnny: There you go, so we’ve not missed a single one, it’s been two a week for, I believe for all those hundred weeks.

[00:09:19] Santiago: Including Christmas, New Years, and all that time when a lot of podcasts tend to take a break. We didn’t take a break, we carried on.

So you found some initial software that we could use.

[00:09:37] Johnny: Yeah, that’s why I was wondering what year it was because it was a time when AI powered tools were being heavily hyped and advertised and, you know, some of them were coming to fruition as being very useful tools.

And it was just around that time I happened to have seen an AI assisted editing software suite advertised, which uses machine learning in that it analyzes the audio and does a very good transcript, a reasonably accurate transcript of what was recorded. And then it has a quality where you can edit the transcript directly and it will be mirrored in how the audio is edited. I don’t know if that’s a standard podcast editing tool, but it’s certainly something I hadn’t seen before.

Like I’d done a bit of recording of music and things 10 years previously and the kind of free software available for that where you’re working just with sound waves and, you know, zooming in on bits to edit is a very different experience from being able to edit a script as written, delete words, even change and replace words with new ones. And this software in that case will use a AI voice trained on the speaker to come up with a replacement word of what you’ve added in. That’s not a tool we turn to very often, but…

[00:11:04] Santiago: I would disagree, I use it quite a lot. But one of the things that I particularly like about that specific aspect of the software is that, yes, you train it with your voice, but I think that the AI voice generator is quite responsible in the sense that it only lets you generate things that are similar to what was said. It won’t allow you to generate a new phrase with the voice of someone else.

So you can tidy up, but you cannot create something that someone else did not say. And that’s a feature that, it can be frustrating at times because if a word is completely missed in the audio for some reason, sometimes it won’t let you create it, and I have spent 20 minutes sometimes trying to find the same word by the same person somewhere else in the episode or in another episode to copy and paste it in. But I think it’s a very sensible limitation.

[00:12:11] Johnny: Yeah, interesting, I didn’t know it had that limitation. ‘Cause I suppose it’s a bit of a grey area where they draw the line, isn’t it? But that makes a lot of sense.

[00:12:19] Santiago: Yeah, and these days there’s a lot of misinformation and there’s a risk of creating, generating content that is supposed to be by someone and they never said that or anything of the sort. So, having that safety net, or that limitations or restriction on what we can and cannot generate, I think suits our responsible AI ideas.

[00:12:49] Johnny: Yeah, absolutely.

I suppose we haven’t yet said what our roles are currently, but yeah, you do most of the editing, the vast majority of the editing, you will go from the raw recording to edited down version.

[00:13:04] Santiago: So it started, we tried to keep costs minimal. We already used Zoom for calls and we continue to use Zoom for recordings. It proved good enough in terms of getting recordings done, get the audio out and put it into this software. Initially I went through Audacity, which is a free open source, I think, editing software.

As you said, you have to handle the sound waves, and it was a huge pain. So I dropped that process completely. I get the raw recordings, upload them to this other software that we use, does a transcript, as you said, it does it fairly well, but it does require careful attention to detail.

Sometimes the labelling of the speaker is off. Every single episode I edit, I have to change the word items for the word IDEMS, which of course we mention a lot being an IDEMS podcast, it’s in every single episode. It would be great if it was better at recognising those common words, I believe there is a glossary that you can use, but I haven’t found it very useful in that regard.

And I go through, and do a first round of editing, trying to make sure that everything is aligned, the speakers are aligned, I try to get rid of stammering or repetition of words to make the conversation feel like it flows a bit better. If you talk to David one-on-one, you’ll realise that the way he speaks is not quite as it’s presented in the podcast. There’s heavy editing in that side.

And when I did that first round of editing, I was very keen from the start that I wanted to have a second set of eyes going through just to make sure that not only the editing was good enough, but also that things that might not be accurate or that might be controversial or that might not be appropriate to say are picked up, and that’s where you come in.

[00:15:36] Johnny: Yeah, so when you’ve done your editing pass, then I’ll pick up the files and do a second pass. Now at that point, I know that someone’s already listened to the full conversation and done most of the kind of functional edits, so it’s quite minimal what’s required. I’ll usually listen to it at 1.5 speed, or sometimes even faster, to just kind of, yeah, do a sense check and everything.

And also, occasionally there’s a noise or something that needs to be removed, in which case I will go back to a program like Audacity that does a bit more of the fine tuning, although that’s been less relevant than it used to be as I think we’ve, yeah, all got a bit more used to not talking over each other and cutting out background noises and things like that.

[00:16:21] Santiago: And then you add the music.

[00:16:24] Johnny: Yeah. Add the jingle at the beginning and the end. And I export the file and add all the description and metadata before publishing it to our feed. The podcast feed technology, RSS is the name of the kind of platform, which is quite old, I think. I think we used to use it for things like streams of news where you could subscribe to an RSS feed of a news outlet and get it sent to you through some app, you wouldn’t necessarily have to go to the news website. You could have an RSS feed that was even like multiple news feeds put together.

But it’s been used for podcasting as long as podcasts have been around. And it means you publish your feed in one place and then podcast listening apps like Apple Podcasts or Spotify, or what was Google Podcast, but is now YouTube music, will all just be pointed at that same feed and they will extract all the information they need and display it in their relative format. So we just have to manage that one output, that we can update and have control over.

[00:17:29] Santiago: And then Lucie comes in. And for accessibility and for all sorts of other reasons we thought it was very important to publish the transcript of each episode in at least one place, I think it only goes into our website where we have all the episodes. And she grabs the transcript well, I think it’s your role, Johnny, you export not only the audio to publish, but also the transcript, and you create a document with the transcript in a shared folder, and then Lucie reviews and makes sure that the English makes sense and that the text is coherent with what is being said.

And that often happens post publishing, and once she’s happy with it, she sends it back to you to put on the website. Is that right?

[00:18:33] Johnny: Yeah, that’s right. And I think she’s good at picking up kind of fact checking issues as well, whether it’s the spelling of someone’s name that’s referenced or things that might have slipped through both of our nets. So it’s, yeah, great to have that third pass over, someone actually paying attention as they listen through.

[00:18:51] Santiago: And I think, three people involved, I tend to spend between two and three hours, occasionally more, on my first pass. How long do you spend on your…?

[00:19:04] Johnny: It depends quite greatly, but usually only something like half an hour from picking up to having it published.

[00:19:14] Santiago: And I believe Lucie is 45 minutes or so. So it’s less than one person day per episode. So it’s not a trivial amount, but it’s not a huge amount of time either. And of course, in terms of the financial aspect of it, the software that we use requires a subscription, which we pay, so there are additional overheads. I’m not sure if the plugin that we use for publishing is free.

[00:19:52] Johnny: it’s free and it’s a plugin for our WordPress website that we already own and manage. So the cost associated that would be just hosting the actual files themselves, which won’t be, it won’t be nothing but it’s as cheap as it could be really.

[00:20:06] Santiago: Yeah. So it is an investment. But I think IDEMS has always struggled to communicate what it does, what its values are, actual principles, well defined principles are, and this was the most effective and efficient way of getting our stories out there and getting David’s stories.

David is a hugely interesting person with loads of different experiences and stories that are or could be valuable for others. So that was the objective and I think we managed to get to 200 episodes. David originally said that he had potentially only a hundred episodes in him. And here we are celebrating our 200th episode and I don’t think we have any plans to stop anytime soon.

[00:21:09] Johnny: No, I think there’s always a lot more stories to tell. There’s plenty of members of the team that haven’t discussed their work and, yeah, taking stock of the 200 mark, I think they’ve already proven to be very useful in terms of kind of seeing the links between things or having somewhere to point people to on a particular topic that otherwise just wouldn’t be documented or wouldn’t be recorded in any way.

Whether or not, you know, people that are interested in IDEMS are the kind of people that listen to podcasts, having the option of them being able to listen to a podcast on a given topic that might be of relevance to them, or having one of our partners be referred to a podcast that explains something that would otherwise be quite difficult to explain, I think has been extremely valuable.

And I think that’s something when looking ahead, currently we have this big back catalog of 200 episodes and they’re available on all podcast platforms. But we’d like to especially draw out those links and make them more useful as a kind of record of IDEMS’ work and IDEMS’ thought processes and things.

So watch this space for that really. But what that probably looks like is a section of the website that’s more easy navigatable between different podcasts on different themes. It would involve adding more kind of categories and tags that again, we’ve discussed, ’cause there’s a lot of interlocking, interrelated themes and guests and areas of IDEMS work where really the only way to represent that complex web is with a kind of graph of these different nodes.

And really the podcast is a kind of representation of that. It is this network of information and data, that having this, yeah having this medium for it has really enabled that to kind of grow out.

[00:23:05] Santiago: Yeah, and to pick up on one point that you mentioned, it has been quite useful in picking up particular episodes that we’ve recorded to send to potential collaborators, potential clients to say, look, this is how we work, this is work that we’ve done related to what you’re doing or what you want from us. And it has been quite useful in that regard. And your classification that you’re discussing will make that job a lot easier. Now that we have 200 episodes, finding the relevant episode can be quite challenging.

But, it’s been a fun activity, it’s been interesting. I’m a mathematician, I never thought I’d be editing podcasts, but I suppose the attention to detail is what allowed me to get into this successfully. And I enjoy doing it and I hope it keeps growing. We’ve been discussing, since episode 100, how to get higher production episodes or a small series with higher production value, maybe include video.

So as you said, watch the space, different things might come in the next a hundred episodes.

[00:24:30] Johnny: Yeah, look forward to talking again at 300.

[00:24:34] Santiago: Well, thank you very much, Johnny.

[00:24:36] Johnny: Thanks Santiago.