Mechanical sympathy is the ability to have an awareness of and empathy for how a machine works, as well as the capacity to adjust your engagement with that machine to help it perform optimally. If you have some mechanical sympathy for data and spreadsheets, you’ll likely get a lot more mileage out of them than if you expect to be able to squirt any old nonsense into them.
I wrote this article about mechanical sympathy for DBW in October 2015, republished here after DBW's sale.
The way I drive, sometimes, you wouldn’t think I believe in mechanical sympathy. But I can assure you I do. Mechanical sympathy is the ability to have an awareness of and empathy for how a machine works, as well as the capacity to adjust your engagement with that machine to help it perform optimally.
If you have some mechanical sympathy for data and spreadsheets, you’ll likely get a lot more mileage out of them than if you expect to be able to squirt any old nonsense into them.
Take rights data, for instance. The Frankfurt Book Fair, and the other book fairs throughout the calendar, are occasions when the technical debt of rights teams around the world comes back to bite them. Gathering together current and accurate information about which rights you hold, which have been exercised, which expire when, which frontlist titles’ rights are contracted and so on, can be a mammoth task and can weeks of preparation.
Would it upset you to know that I produce my rights guide the night before Frankfurt, by clicking a single button? There are a number of steps to get to this happy state of having unambiguous rights information in one place, synchronized with editorial’s and legal’s data, and you don’t need one particular system to do it. But you do need mechanical empathy and a rigorous business workflow in place to enforce it.
The key is to store data in a way that computers can understand. If you do this, it means that you can query your data programmatically, and that computers will be able to ingest it properly. That gives you a lot of flexibility around which system you use to manage your data, now and in the future. And if computers can understand it, it’s actually more likely that humans can understand it, too. Computers hate ambiguity.
Take these two examples of how to store data. Which do you think can easily be imported into a database, an InDesign template or a new publishing system?
The first spreadsheet has no end of technical debt being stored up:
- No validation of the title. There’s nothing to stop someone typing in numerous variations of the same data: “Brave New World,” “Brave new world,” “ Brave New World” with a leading space. Use filtered drop-downs to ensure there’s only one valid version of each title. Relational database management systems do this through associations between tables; you can do it in a spreadsheet by setting up list validations.
- More than one sort of data in a cell. By putting “France” and “Spain” in the same cell, it’s hard to associate other data properly—which is why the next column has to go on two rows.
- Imprecise categorisations. When we say “Rights sold,” do we mean the language, the script, the territory or what? France is a territory, German is a language, Simplified Chinese is a combination of a language and a script.
- Conflicting types of data in the same column. The Advance column mixes up currency amounts, descriptions of payment terms and payment statuses. How could a computer figure out what the advance amount was, the currency, or the payment date, from the string “Paid?”
- Inconsistent ways of formatting data. The last column presents date information three different ways.
- Multiple possible values for one piece of data. If you’re going to have one line per rights sale of a title, you might be tempted to have a column showing which rights you own. But that would mean putting, say, “World” on multiple lines. If one of those lines says “Commonwealth” and the rest say “World,” which one are you going to use as the right answer?
Using colour-coding, indentations and font sizes in spreadsheets to communicate categorisations of data are more ways that make it very difficult for a computer to figure out what’s true. The reality, however, is that, while you might feel that this is “your” spreadsheet, it’s really not. The data belongs to your company, and you are obliged to store it in such a way that other people and other systems can understand it.
Store your data properly, and you’ll be able to get the most out of whatever systems you want to use. Then you can spend your time preparing for the substance of meetings at the book fairs instead of burning the midnight oil hand-writing a rights guide that’ll be out of date before you’ve even finished.