The business case for culling software: the payback math

Nobody has to approve your subscription. You still need a reason you will believe in November, when the season is over and the charge lands anyway. This is the case laid out the way you would put it to a boss: what it costs a year, what an hour of yours is worth, how many hours it has to give back before it pays for itself, and the conditions under which the honest answer is no.

Write it as though you had to ask someone

In a studio with a manager, buying a new piece of software means writing something down. What it costs. What it saves. When it pays for itself. What would have to be true for this to be a mistake.

Solo photographers skip that step, for the excellent reason that there is nobody to hand the memo to. So the decision gets made in one of two bad ways: on a Sunday night at frame 2,400, when anything looks like a good idea, or never, because it never feels urgent on a Tuesday.

Write it anyway. Four lines fit on one page, and the point is not rigor for its own sake. A decision you can defend to yourself is a decision you stop relitigating every time the renewal email arrives. Here are the four lines, in the order to fill them.

Line one: the annual cost, never the monthly one

Monthly pricing exists to make software feel small. Multiply by twelve before you look at anything. Here is what the field costs a year at the entry level, from each vendor's own pricing page.

Prices checked August 2026 from each vendor's own pricing page.

Two notes before you pick a row. Imagen's culling is an add-on rather than a standalone product, so the true annual figure is $144 plus whatever the editing plan costs, and editing is the core of that product. Photo Mechanic is a different kind of purchase: a famously fast browser and ingest tool with no AI selection in it. It makes you faster, it does not decide anything for you. Its perpetual license at $299 also changes the payback math completely, because it stops charging you.

For scale, The Knot's 2025 study of 10,474 US couples puts the average wedding photographer at $3,000 a booking, with quartiles at $1,500, $2,800 and $4,700. Every line in that table is a fraction of one job. That is not an argument on its own, just the size of the first number, and a small number still fails a case when it buys nothing.

One disclosure, since we build a tool in this space. Kepla's picking app for iPhone, iPad and Mac is still being built, so it is not a row in that table yet. Founding members get it free while we build. The booking page is live and free today.

Line two: what one hour of yours is worth

You cannot calculate a payback without a value per hour, and this is where most photographers wave their hands. There are three defensible methods. Pick one, write the number down, and then use the same one for the rest of the memo.

Replacement cost

What you would pay a competent freelancer to make the first pass for you. This is usually the lowest of the three and the easiest to defend, because it is a price somebody would actually accept.

Displacement value

What the hour would earn if you spent it on the next best thing: shooting, selling, answering an inquiry that is sitting in your inbox. This is the highest number and the least honest, unless you would genuinely fill the recovered hour with that work. If the hour goes to sleep or to your family, that is a real benefit, but it is not revenue.

Blended job rate

Your package price divided by every hour the job touches, from the first email to the delivered gallery. This is the number we walked through in the hidden cost of slow culling, and it is the most realistic picture of what your time is currently earning.

Whichever you choose: my hour is worth $______. Mixing the three is how a business case quietly becomes a rationalization, so pick and stick. If this exercise makes you want to reprice the work rather than resource it, that is a legitimate outcome, and pricing for your editing time is where to take it.

Line three: the breakeven, in hours

Now the math, which is one division.

• Annual cost of the tool: $______
• Value of one of your hours, from line two: $______
• Divide the first by the second: ______ hours it must save in a year
• Shoots you do in a year: ______
• Divide again: ______ hours it must save per shoot

Take a real figure to see the shape of it. Aftershoot Select is $120 a year on annual billing. Divide $120 by whatever you wrote in line two and you get the number of hours the tool has to give back across twelve months to break even. Divide that by your shoot count and you get the per shoot requirement, which for most working photographers comes out in minutes rather than hours.

That is why so few of these cases fail on the arithmetic, and it is also why the arithmetic alone should not convince you. Two traps sit underneath it.

Gross hours are not net hours. The saving is not the whole time you used to spend. Subtract the analysis time while the software works through the folder, the review time going through what it proposed, and the correction time putting back the frames it did not understand. What is left is the real saving, and it is smaller than the marketing. We took that apart properly, using only the one survey worth citing, in how much time AI culling actually saves.

When the hours fall matters as much as how many. Three hours saved across a quiet January is worth less than the one hour that lets a gallery go out in week two instead of week seven. That does not appear in the division above, which is a limitation of the division, not a reason to inflate it.

Line four: the conditions that make this a no

A business case with no failure conditions is an advertisement. Write these down before you buy, so that in six months you can check them rather than argue.

• Volume too low. If you shoot a handful of small sessions a year, the hours saved may never clear even $120, and no amount of enthusiasm changes that. Run line three before you assume otherwise.
• Genre mismatch. These tools check for missed focus, closed eyes and near identical frames. If you shoot work where deliberate motion, unconventional focus or an odd expression is the point, the software is grading you against criteria you broke on purpose. Conceptual, fine art and heavily stylized documentary work all sit here.
• Files that cannot leave your machine. Imagen's culling runs in a web interface with photos uploaded first, which is a real constraint if you have a contract, a client policy or an ordinary upload speed. Aftershoot, Narrative and FilterPixel process locally.
• Seasonality. A subscription bills in February as cheerfully as in August. Aftershoot publishes the ability to pause for three months a year. Check what each vendor allows before assuming twelve months of value for twelve months of cost.
• Temperament. If you know you will re-check every frame yourself regardless, you have bought a second opinion rather than a time saving. That is a legitimate thing to buy. Just price it as what it is, and do not put hours in the savings column that you are not going to take.

If more than one of those lands, the honest conclusion may be that the answer is not yet. Is AI culling right for you walks the same decision from the other end, by volume, genre and hardware.

The evidence to collect before you sign anything

Every serious tool in this category gives you a free look, and most of them do not ask for a card. Aftershoot runs a 30 day trial with no card required. Narrative offers a free full trial of its top Ultra tier, also with no card. FilterPixel gives you four complete projects with unlimited photos. Imagen includes two free culling projects. Photo Mechanic runs 30 days, one per customer.

That is more than enough to replace the guessing in line three with a measurement. Here is the test worth running.

1. Take a shoot you culled by hand some months ago, ideally one you have half forgotten.
2. Run it through the trial. Time it on the wall clock, from opening the folder to having a shortlist you would actually send, not the processing time the software reports.
3. Then look only at the frames where the tool disagreed with you. Agreements teach you nothing. If it left out shots you kept for emotional reasons, that is what it will cost you. If it kept frames you ruled out on technical grounds, that is how much reviewing you will be doing forever.

Two numbers go back into the memo: hours the pass took by hand, and hours it took with the tool. The difference, multiplied by your shoots per year, is your savings line, measured rather than claimed. That number is worth more than every vendor benchmark put together. If you are still choosing which trials to run, how to choose culling software narrows the field to two or three first.

The line most cases miss: the rest of the stack

Culling software never arrives alone. It joins a booking tool, a gallery host, a contract service, a scheduler and a CRM, and each of those was approved the same way this one is about to be: individually, on a night when it seemed reasonable, without anyone laying the twelve monthly charges next to each other.

Before you approve line one, put the whole stack in the cost calculator and look at the annual total in one place. Either the culling subscription looks trivially affordable against everything else, which settles the question, or you find two tools whose jobs overlap and the real decision turns out to be a swap rather than an addition. We went through what a full stack typically runs in what a photography tool stack actually costs.

Worth naming plainly: on our own numbers, replacing a typical stack of separate booking, gallery, scheduling, signature and culling tools works out at roughly $876 a year saved, and the calculator shows the working rather than asking you to take our word for it. Kepla's booking page is the part of that which is live and free today. The picking app is still being built.

The whole memo on one page

Six lines. Fill the middle column and keep it somewhere you will find it at renewal.

The decision rule is one comparison. If your measured saving per shoot is comfortably larger than the per shoot requirement, approve it, and the gap between them is your margin of safety. If the two are close, it is a no this year, and you revisit it when your volume rises or your hour becomes more valuable.

Then add one sentence the table cannot hold. Some of what you are buying does not fit in an hourly rate: galleries going out while the client is still excited, and the quality of your judgment at frame 2,400 compared with frame 40. Those belong in the case as a stated reason, not as a number invented to make the total look better. A memo that admits which parts are unmeasured is one you can still trust in November.

FAQ