Last month I timed myself building a lower third from scratch for one show. Fourteen minutes. Then I timed myself dropping our template version of the same lower third onto a timeline and retyping the guest name. Nineteen seconds. That gap is the whole argument, and most people get the wrong lesson from it.
The wrong lesson is "templates win, use templates for everything". The right lesson is that fourteen minutes is a price, and a price is only bad if you are paying it for the wrong thing. I pay it maybe twice per show, once, at the start. After that I never pay it again.
So here is my actual position, and it is going to annoy both camps: almost every graphic in a podcast episode should be a template, and the two or three that decide whether people recognise your show should never be.
What people actually mean when they say "custom"
When a host tells me they want custom graphics, nine times out of ten they mean they want graphics that look like their show and not like a stock pack. That is a branding request, not a production request. It gets solved once, in a design pass, and then it lives inside templates forever.
The other one time out of ten, they mean they want a graphic that responds to what is happening in the episode. A chart that appears when the guest cites a number. A map when they mention a city. A text callout that reacts to a specific joke. That is a genuinely custom, per episode, per moment build, and it cannot be templated because the content is different every time.
Those two requests get the same word and they cost wildly different amounts. Confusing them is why so many podcast graphics conversations go sideways.
The 90/10 split I actually run
Look at a real episode and count the graphics. A typical one we cut has something like:
- An intro or title card
- A guest name lower third, sometimes two if there are two guests
- A show logo bug in a corner
- Chapter or topic cards between segments
- Two or three text callouts for quotes
- An outro card with the subscribe ask
- Maybe an ad or sponsor bumper
That is around ten to twelve graphic events. Of those, maybe one is genuinely episode specific. The rest are the same shapes with different words in them. Building all twelve fresh every week is not craft, it is just an expensive habit.
Templates are not the opposite of custom. A good template is custom design that you only had to pay for once.
The 10% that stays custom is worth guarding. It is usually the intro, and it is usually whatever visual thing only your show does. If the whole episode is templated including the opener, the show starts to feel like a format rather than a show. You want the first eight seconds to be unmistakable.
Where templates quietly fail
I am not going to pretend the template side is free. Three failure modes come up constantly.
The first is text overflow. Someone builds a lower third around a four letter first name and then a guest turns up called something with eighteen characters and a job title that runs to two lines. The box does not stretch, the text shrinks to something illegible, and nobody notices until it is published. Build every text template around the longest plausible string, not the shortest.
The second is aspect ratio. A lower third that sits beautifully in a 16:9 frame ends up sitting across someone's mouth in a 9:16 crop. If your show makes vertical clips at all, and it should, the template needs a vertical variant designed on purpose, not a scaled down copy. I have seen more branding undone by a squashed logo in a Reel than by any amount of bad design.
The third is drift. Six months in, someone tweaks the colour on one card, someone else changes the font weight on another, and you now have four slightly different versions of your own brand in circulation. Templates need an owner and a single source file, or they rot.
Multi-cam changes the maths
On a single camera show, graphics are the main visual variety you have, so they carry more weight and I lean toward spending real design time on them. On a two to four camera show, the cutting itself provides the variety. Angle changes do the work that a busy graphic would otherwise have to do.
That is why our multi-cam editing work usually ends up with fewer graphics than the single camera equivalent, not more. Hosts sometimes expect the opposite. They assume the bigger production tier means more stuff on screen. It usually means less stuff, placed better.
This is also where the recording setup starts to matter for graphics. If you are recording remotely, the way your files come out determines what a graphic can even sit on top of, which is most of what I get into in the piece on how double ender recording works. Clean isolated tracks mean clean frames. Compressed single stream recordings mean you are placing crisp vector graphics on top of mush, and the mismatch is visible.
The handoff problem nobody plans for
Here is the practical reason I push templates harder than my design instincts want me to: templates survive handoffs and custom builds do not.
A custom graphic lives in one project file, on one machine, made by one person, in one version of one app. The moment that person is on holiday or you switch editors, it is gone. You get a rebuild, and the rebuild is never quite the same. A template pack lives as an asset with rules attached, and any competent editor can pick it up on day one.
We learned this the hard way and it is a big part of why I now insist on structured asset folders, which I wrote about in more detail in the file handoff post. If your graphics only exist inside a project file, you do not have graphics. You have a hostage situation.
The same logic applies when you are comparing an individual freelancer to a studio setup. It is not really about skill level, it is about whether the system outlives one person, which is the honest version of the freelancer versus agency comparison I keep having with hosts.
What I do when the raw footage is rough
There is one case where I flip my own rule and go custom heavier: bad source material.
When an episode has a camera that drifted out of focus, or a guest whose webcam died halfway through, graphics stop being decoration and become repair. A full screen quote card is a legitimate way to cover forty seconds of unusable video while the audio carries the moment. That is not a template decision, that is a per episode judgement call about what is broken and what will hide it.
I go deeper into that whole toolkit in rescuing a badly recorded episode, but the short version is that repair graphics are custom by definition, because damage is custom by definition. A template cannot know where the hole is. A human watching the whole episode can, which is exactly why nothing we do gets auto cut by silence detection. Silence detection cannot see that a shot is unusable. It only hears gaps.
Thumbnails are a separate discipline, stop templating them
One thing I will argue hard about. Episode thumbnails should not be templated in the lazy way most shows template them, which is to say the same layout with a new face pasted in and the guest name in the same corner every week.
In an episode, templates work because the viewer has already decided to watch. Consistency is comfort. A thumbnail has the opposite job. It is competing for a click against everything else in a feed, and consistency there just means blending into your own back catalogue. A recognisable frame or colour system is good. An identical layout is not.
That is a genuinely different skill from episode editing, which is why our sister studio ClickTheory handles thumbnails separately at 8 for $250 rather than us treating them as one more card in the template pack. Different job, different brain.
How I would build a graphics system from scratch
If I were starting a show tomorrow, here is the order I would do it in.
- Pick two fonts and three colours. Stop. Do not pick five colours.
- Build the intro as the one genuinely custom piece. Give it real time.
- Template everything else off that intro: lower third, chapter card, quote card, outro.
- Build a vertical variant of every template on day one, not in month four.
- Save the whole pack as its own folder with a plain text file describing the rules.
- Leave one slot per episode open for a custom graphic if the content needs it. Usually it will not. Occasionally it will save the segment.
That is maybe a day of work up front and then close to zero per week. Compare that to rebuilding the same lower third fifty two times a year.
The graphics system also feeds everything downstream. Clips inherit it, the show page inherits it, and if you are running any kind of email push off each episode, the visual language carries into that too, which is half of what makes a newsletter loop feel like part of the show rather than a separate marketing thing. Templates for the 90%, custom for the 10% that carries your identity or repairs a real problem. Build the pack once, own it as an asset, keep a vertical version, and let a human decide each week whether this particular episode actually needs something new.
If you want a second opinion on which of your graphics are pulling weight and which are just habit, send me an episode at samy@pixelsamy.com and I will tell you what I would cut. No pitch attached. It is a fifteen minute look and it usually saves people more time than it costs me.