The real storage cost of overshooting, worked out

Pressing the shutter costs nothing, which is why the bill arrives years later and in a different currency: drives, backup windows, migration days, and an archive nobody can search. This is the arithmetic, laid out so you can fill in your own numbers. Six figures you already know, four multiplications, and one honest answer about which extra frames are worth their keep.

The whole thing is one line

Every argument about overshooting can be settled with one calculation:

Five year footprint = average RAW size x frames per job x jobs per year x copies you keep x years you keep them.

That is it. The reason the conversation usually goes nowhere is that people argue about the first two terms and quietly ignore the last three, which are the ones that multiply.

You need six numbers, and five of them are already on your drives:

• The size of one of your own RAW files. Open a folder from last month and look. Resolution, compression setting and format move this enormously, so a number from a blog is worse than useless here.
• Frames captured on a typical job, not your best guess. Your last three shoots will tell you.
• Jobs per year.
• Copies you keep of each job.
• Years you keep them.
• An allowance for everything that is not a RAW file: previews, catalogs, sidecars, exports, video clips, and free space so the drives keep working.

The example below uses 45 MB files and 4,000 frames because the numbers stay legible. Substitute yours. If you want the related ratio, how many of those frames end up delivered, that is keep rate, and it is worth tracking alongside this.

Step one, one job

Average RAW size times frames captured.

45 MB x 4,000 frames = 180,000 MB, which is 180 GB for one wedding in the decimal terms drive manufacturers use. Add previews, the catalog, sidecars and the delivered exports and the working job is comfortably north of 200 GB before anything is backed up.

Now the interesting part, the marginal frame. The gap between a 3,000 frame wedding and a 4,000 frame wedding is not abstract:

1,000 extra frames x 45 MB = 45 GB, per wedding, before you make a single copy.

Forty five gigabytes sounds like nothing. Hold on to it, because it is about to be multiplied three times.

Step two, multiply by the copies

One copy is not an archive, it is a single point of failure with your business inside it. The working rule is three copies, on two kinds of storage, one of them off site, and every one of those copies pays full price for every frame.

So one wedding is not 180 GB, it is 540 GB of capacity plus overhead. And those 1,000 extra frames are no longer 45 GB, they are 135 GB per wedding, sitting on three separate things you have to buy, verify and eventually replace. The copies are not waste, they are the difference between a dead drive and a dead business. They are also the reason volume compounds. The structure itself, which drive does what and how you verify a copy, is in the 3-2-1 backup workflow.

Step three, multiply by the season and the years

Twenty weddings a year at 4,000 frames is 80,000 frames, or 3.6 TB of RAW originals annually. With three copies that is 10.8 TB a year. Keep five years of work and you are looking after 54 TB of capacity, plus portraits, commercial jobs, personal files and enough free space that nothing is running full.

Fill in the right hand column with your own figures. It takes about five minutes and it is the most useful five minutes in this article.

You do not buy 54 TB on a Monday. Capacity is added over years, which is exactly why the forecast is worth having: it tells you whether "keep everything forever" is a plan you have funded or a sentence you have said. Run the same table with 1,000 fewer frames per wedding and watch what falls out of line I.

What that capacity costs in money you can look up

Hardware prices move too fast to quote usefully, so price drives yourself on the day you buy. Cloud is published and stable enough to anchor on.

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

Put the arithmetic through the B2 rate and the picture gets concrete. One wedding at 180 GB is 0.18 TB, so a single off site copy of it is about $1.25 a month, roughly $75 over five years. Set against The Knot's 2025 average of $3,000 per US wedding, no single job is ever going to justify clearing files in a hurry.

The compounding is where it bites. Five years of twenty weddings is 18 TB of originals, and one cloud copy of that is about $125 a month, near enough $1,500 a year, and still climbing. And those 1,000 extra frames per wedding, alone, come to about 4.5 TB after five years, roughly $31 a month for one cloud copy of frames you already decided against. Two of the three copies are still hardware you own. Which architecture makes sense at your volume, hot versus cold, drives versus cloud, is covered in cloud storage for photographers, and if you want to see this next to the rest of your software bill, the cost calculator does that job.

Read the model of each product, not just the price. Object storage bills for what you keep, so it grows with the archive. A backup subscription bills flat but is designed to mirror what is currently attached to your machine, which is a different promise from long term archival. Read the terms on drives you only plug in occasionally.

The costs that never show up on an invoice

Drive money is the smallest part of this. The rest is paid in hours.

• Ingest and verification. Every frame is copied and checksummed at least twice on the night of the shoot, when you are least interested in waiting.
• Preview building. A bigger catalog is a slower catalog, on every machine that opens it, forever.
• Backup windows. Uploading a wedding on a domestic connection can occupy the same bandwidth your client gallery is trying to use.
• Migration days. Drives age out. Moving many terabytes and verifying it can take a working day you did not plan for, and it comes around every few years.
• Retrieval. The bigger and messier the archive, the longer it takes to answer "can you find one photo of my dad from 2022," which is the moment the archive either earns its cost or does not.
• Sorting time. More frames means a longer first pass. A UK survey of more than 300 wedding photographers, run in 2020 before AI tools were common, put culling at 11% of working time and shooting at 4%. The frames you add on the day are paid for at your desk.

None of these appear when you compare drive prices, and together they are usually larger than the storage line itself.

Which overshooting protects you, and which is just hesitation

A high frame count is not the problem. Overshooting means capturing frames that do not raise your chance of a better or safer result, and the distinction is easy to see once you look for it.

Frames that earn their storage: the burst through the first kiss, a child who is about to do something unrepeatable, the safety frame on a group of thirty where somebody always blinks, action at unpredictable speed, a bracket you genuinely need. This is insurance on moments that happen once, and the case for it is made in safety shots and overshooting.

Frames that do not: thirty exposures of a static posed portrait where nothing changed, repeated test frames nobody flags afterward, every dance floor moment turned into a twenty frame sequence out of habit, and long bursts fired because the decision felt hard in the moment. That last one is the honest category. Hesitation at the camera becomes comparison work at the desk, then triple storage for five years.

Audit yourself rather than guess. Take three finished jobs, sort by capture time, and look for the clusters. You will recognize your own pattern within a few minutes, and it is usually one or two specific situations rather than a general habit. Fixing the technique for long sequences is covered in culling high frame rate bursts. Do not change how you shoot mid season on a hunch, and never shoot timidly on the days that cannot be repeated. Coverage is the job.

Two levers, and you need both

The first lever is at the camera and it only works on the low value clusters you just identified. Pause after you have the expression. Use a short deliberate burst for movement rather than a default one. Check the settings that generate accidental frames. That is it, and it is worth maybe a fifth of your volume on a good day.

The second lever is retention, and it is the bigger one, because it decides how long everything you did capture stays multiplied by three. Separate the archive of originals from the permanent record of delivered work: final edits, album files and selected originals are small and worth keeping for a long time, while every unpicked original does not need the same treatment forever. What that policy looks like, including which frames are genuinely safe to clear and the gates to pass first, is in should you keep rejected RAW files. The short version is that keep is the right default and the exception is narrow.

A tighter first pass helps both levers, because it is what tells you which clusters were pointless and which alternates are worth carrying. Kepla clears the obvious misses and hands you the clean frames, and it never deletes, moves or renames a single file: it marks what it picked, and everything else stays exactly where it is. The apps for iPhone, iPad and Mac are still being built, and the booking page is live and free today.

Do the table once a year. Storage stops being an emergency purchase and becomes an ordinary line item you already knew was coming.

FAQ