Do you always start over a new font from scratch?

I always start from scratch. With an idea to create something new!!
The exceptions being revivals of 1900 fonts, scanned and digitized.

Why reuse or repurpose existing fonts? There are tons and tons of fonts already existing. And for all kinds of use, text, advertising, titles, etc. All kinds of styles too.

If I have a look at my 1981 Letraset catalogue, there were already hundreds and hundreds of fonts available. And most have been professionnaly digitized.

IMHO; nobody really needs an umpteenth repurposing of Caslonesque, Helveticesque or Bodoniesque fonts.

Be original, Filip !

I forgot to mention that, like Mark, I also sometimes start from one of my previous fonts, with a new idea. (I'm not a professionnal, I must add.)

I’ve never considered starting with an existing typeface. I don’t think it would work well to try and adapt one design to fit another.

I don’t have a single way of working, so typically figure out for an individual project whether it makes most sense to start from scratch or from outlines from a previous project. The latter may save me time, and I don’t feel it restricts or directs the design direction, because they’re just points and curves and I can drag them into any new configuration. I always have a shape in mind, regardless of whether I start from scratch or with an outline to manipulate.

I’ve also done jobs in which starting from another design was part of the brief. The Sanskrit Text font that Fiona Ross and I made for Microsoft was directly based on Monotype Devanagari and Bembo—to which Microsoft have ‘ownership-like rights’—and we started from the outlines of those fonts.

What might be the reason for a client to order a font that is not so original? Why would they order a font that "already exist" instead of purchasing it directly?

1. Because it may be less expensive to commission a clone for a one-time payment, rather than an enterprise-wide renewable licence for an existing typeface.

2. Because they have developed their branding using an existing font as a placeholder, but would now like to fine-tune the concept with a custom design.

@Alex Visi It’s not a difference in the term (time period) of the license per se, but in the term of exclusivity. Lifetime license either way, usually.

It depends on who the client is, and how smart (=thinking about Net Present Value in accounting/economic terms) the foundry is.

For a client like an art director of a magazine (or many kinds of media), an exclusivity period of two years or five years at the very most might be enough to extract nearly all the value they get from exclusivity. Part of this is because the art director is likely to have moved on. Perhaps part is because after a few years nobody is paying attention to the unique typeface any more. Whatever the reasons, foundries have discovered that many clients are open to this!

On the other hand for the foundry, aside from questions of popularity that may well vary over time, it is otherwise a normal Net Present Value calculation where you discount for inflation/alternative investment at say 5% per year. (Or something.)

So, given that the client’s value for exclusivity is often (for many common clients) slanted much more heavily to the near-term than that of a rational economic-minded foundry, in such cases it is often in BOTH parties’ interest to limit the period of exclusivity. The foundry can then afford to charge the client less—enough less that it is a “better deal” in terms of their valuation of near-term exclusivity. And the foundry can “get the design back” in a few years and offer it as a regular retail product or whatever they wish to do with it, and also come out ahead.

Alex Visi said:

Does anybody know how common life-time licenses are? Some foundries limit the license to a few years period even for custom fonts, after which the fonts get published through the shop, instead of giving away all the rights forever.

I recently did one for a multimedia company; since they were willing to pay me a significant amount, I had no problem handing over the rights.  The thing is that I made it clear that since it was based on an existing font I did, I would ultimately include some - not all, some - of the things I did in the custom font into the parent font, they had no issue with it.

Ironically, the company just went belly-up recently, so I have no idea who owns the font now.  I'll probably fork the changes into the parent font eventually.

My understanding is that San Francisco was developed in-house at Apple, not by an outside designer. If any outside designers were involved, they haven't spoken publicly about it that I'm aware of.

I never base my new typefaces on previous ones. I feel like it prohibits my creativity too much and locks me into a certain framework or way of thinking. Would always advocate for this method, rather than a more iterative one.