The first time I built a Photoshop action, I was 26, working a late shift at a studio in the West Loop, and I had just spent four hours manually resizing and sharpening 80 product images for a catalog client. Same crop. Same output size. Same sharpening radius. Eighty times. By image 60, I wasn’t even seeing the photos anymore, just clicking through the motions like a machine pretending to be a person.

I went home, watched a single YouTube tutorial on actions, came back the next morning, and automated the whole thing in about 45 minutes. That was my introduction to what presets and actions can actually do when you treat them as infrastructure rather than conveniences. Fifteen years later, I run a post-production consultancy out of Chicago, and that same philosophy, building systems instead of solving problems one at a time, is at the core of everything I do for ad agencies and e-commerce clients.

Here’s the problem I keep seeing: most photographers and retouchers use presets the way they use bookmarks. They save one when something looks nice, forget about it, save another, end up with 200 untitled presets and no coherent system. The presets become clutter instead of tools.

The Difference Between a Preset and a Workflow Component

A standalone preset adjusts your image. A workflow component adjusts your image in the right context, at the right stage, for the right output. That distinction sounds abstract until you realize that a sharpening preset applied before you resize is doing completely different work than the same preset applied after.

Photoshop’s sharpening algorithms, Unsharp Mask and Smart Sharpen both included, are resolution-dependent. A radius of 1.2 pixels at 300 DPI is not the same as a radius of 1.2 pixels at 72 DPI. If your output sharpening preset doesn’t account for the resolution of the file it’s touching, you’re getting inconsistent results and probably blaming yourself for it. You should be blaming your preset architecture.

This is why I build presets in stages: a base correction preset, a mid-process preset tied to a specific output size, and a final export preset. They only work correctly in sequence, and that sequence is locked into an action. The preset is the setting. The action is the context.

How I Structure Custom Presets for E-Commerce Batches

For a typical e-commerce client running 500 SKUs, my system uses three action sets, each color-coded in the Actions panel (yes, I color-code my actions, everything else in my life too if I’m honest). Red for destructive or irreversible steps. Green for export. Yellow for anything that requires a manual check before the action continues.

The base preset layer handles white balance normalization, exposure correction in the range of plus or minus 0.3 EV, and a calibration shift specific to the studio’s lighting rig. That last part matters. Every studio has a color cast. If you’re not baking a compensation for your specific lighting into your base preset, you’re correcting for it manually on every single image.

The mid-process preset handles background cleanup parameters and a selective sharpening layer using a high-pass filter at 3 pixels radius, blended in Overlay mode at 65% opacity. I land on that number after testing across matte, glossy, and textile product surfaces. It holds up across all three without oversharpening edges on fabric or creating halos on glass.

The export preset resizes to 2000 by 2000 pixels at 72 DPI, runs a final output sharpening pass at Unsharp Mask settings of Amount 85, Radius 0.8, Threshold 3, then saves as JPEG at Quality 9 to a timestamped subfolder. Quality 9 rather than 10 cuts file size by roughly 40% with no visible difference at web resolution. On a 500-image batch, that can mean the difference between a 4GB delivery folder and a 2.4GB one. Clients notice when transfers are faster.

The Variable Nobody Talks About: Preset Degradation Over Time

Here is something I’ve learned the hard way: presets degrade. Not technically, but contextually. A color grading preset you built for a client’s product line in 2021 was calibrated to their brand guidelines in 2021. If their guidelines shifted, if their packaging changed colors, if they rebranded, your preset is now fighting against the current brief without telling you.

I audit every recurring client’s preset stack once a quarter. It takes about 90 minutes per client, and I track those sessions on the same spreadsheet I use to log total time saved from automations overall. The time spent auditing is almost always paid back within two batch jobs, because catching a drift in a preset before you’ve processed 200 images is significantly better than catching it after.

Building Presets That Other People Can Actually Use

If you’re running a team or handing off files to a freelancer, your preset system needs documentation the same way your code needs comments. I keep a one-page PDF inside every client’s preset folder that explains what each preset is for, what file state it expects as input (flattened vs. layered, resolution, color profile), and what it’s not designed to handle.

This sounds like extra work. It is extra work, upfront. But the first time a junior retoucher applies your base correction preset to an already-color-graded file and you have to explain why the skin tones look green, you will write the documentation.

The preset file itself should be named with version numbers and dates. Not “SharpenFinal.” “Sharpen-Output-v3-2024-09.” Future you is a different person with different problems, and future you deserves clarity.

One Number That Changed How I Think About This

I keep a running total of hours saved by my action and preset systems. Right now it sits at over 2,400 hours. That number isn’t there to impress anyone. It’s there because when I’m spending a Saturday afternoon building a new preset structure instead of billing hours, I need a concrete reason to believe the investment is worth it. Two thousand four hundred hours is the reason. That’s roughly 300 eight-hour workdays returned to me because I treated automation as infrastructure from the start.

Build your presets like you’re building a pipeline, not decorating a room. The goal isn’t a collection of looks, it’s a system where every step knows exactly where it sits, what came before it, and what comes next.