When lost client photos urgent situations land, the first decision is not technical. It is which card to work on and what the client hears while you work. Triage by recoverability multiplied by client value, not by client value alone. The card you care about most may be the one that has already been overwritten, and spending four hours on it while an untouched card waits is the expensive mistake.
This page sets out a working triage order, the imaging step that precedes any scan, when the client has to be called, and how partial deliverables hold a relationship together.
The First Ninety Minutes: Preserve Before You Analyze
Under deadline pressure the instinct is to open recovery software immediately. The correct first move is to stop all writing to every card in the set, then preserve what exists.
That means three things in order: stop using every card involved, including the ones that look fine; note or photograph the state of each card so you can reconstruct what was done to it and when; and make a byte-for-byte image of every card before scanning any of them.
Imaging matters because a scan is not free. Mounting a card lets the operating system write its own metadata, and any tool that “fixes” a volume writes to it. Those writes consume free blocks and can overwrite the very clusters you are trying to recover. With images in hand you can scan the same data repeatedly, with several tools, at no further risk. The procedure is in imaging a card before recovery.
Triage Order: Rank by Recoverability, Not by Importance
The emotional ranking of cards and the technical ranking are different, and following the emotional one is how a recoverable shoot gets ruined.
Score each card on client value, remaining intactness, and scan time. The middle axis is the one people skip, and it decides the outcome.
| Card state | Data still intact? | Priority |
|---|---|---|
| Pulled and set aside, never powered since the loss | High probability | Highest; image and scan first |
| Removed normally, but the camera kept recording | Moderate; some clusters reused | Second, after imaging |
| Formatted in camera, then used for another shoot | Low for the earliest files | Third; set expectations low |
| Formatted and refilled with new footage | Nearly zero in overwritten areas | Last, or skip entirely |
| Handled by a shop that ran a repair utility | Unknown and degraded | After intact cards, from an image only |
Two conclusions follow. An untouched card is the highest-probability target, so work on it before the client conversation consumes your evening. And a card that was formatted and refilled is usually beyond software no matter how important it is; overwritten data is not recoverable by any tool, so continuing to spend the client’s timeline on it costs you the cards that can still be saved. If the most valuable card is also the most damaged, put it down and work the second one.
Do Not Do These Because You Are in a Hurry
Deadlines produce a specific set of destructive shortcuts, each of which has cost a photographer a shoot that was otherwise recoverable.
Do not format anything to “start clean.” Do not let the camera or drone write to the affected card again, including test frames. Do not accept an in-camera repair prompt on a card you have not imaged. Do not hand the card to a colleague who wants to try a different tool. Never save recovered files back to the source card. The complete list, with the mechanism behind each, is in the mistakes that end recoveries.
The costliest single mistake under pressure is running a utility that offers to repair the file system. It writes a new allocation table over the old one, converting a readable but unmounted card into a mounted card with no files. That is the moment many projects stop being recoverable at all.
When to Call the Client
Most guides say to call immediately. That is not quite right, and the distinction is commercial rather than technical.
Call as soon as you have a diagnosis and a plan, which is usually ninety minutes to a few hours of work. A call made at minute five transfers panic without information, and panicked clients act in ways that make recovery harder, such as demanding a second opinion that adds another set of writes to the card.
The line you must not cross is the client’s own decision point.
| Situation | When the client hears from you |
|---|---|
| Client’s decisions do not change for a week or more | As soon as you have a diagnosis and a timeline |
| Client is choosing today between you and a reshoot | Before they decide, even with incomplete information |
| A broadcast, print, or exhibition date is fixed | Immediately, with the facts you have and a next-update time |
| Recovery is complete and verified | After delivery; no reason to raise alarm |
When you call, lead with what is intact. “Six of the eight cards are fully readable and already copied” is true in most multi-card losses and gives the client something to hold. Then state what is at risk, what you are doing, and exactly when you will next update them. Never give a percentage you cannot support. A recovery rate spoken under pressure becomes the expectation the client measures you against.
Deliver Partial Results as They Land
The instinct under pressure is to withhold everything until the full recovery is done. Deliver earlier instead, because the client’s anxiety is mostly about whether the images exist at all.
Proxy and thumbnail files are separate objects on the card, so a truncated master does not imply they are gone. A proxy sequence is not a deliverable, but it is proof of coverage and it lets the client choose selects.
Embedded previews inside RAW files are the most underused deadline tool available. A RAW container stores a full-size JPEG preview near the start of the file. When a RAW will not open because its container or maker note is damaged, that preview is frequently still complete, and extracting it yields an image good enough for web, a contact sheet, or the client’s social posts. The method is in pulling the JPEG out of a RAW file.
Verified full-resolution files should go out as scans complete rather than waiting for the whole set. A practical schedule for a wedding-scale loss is proxies and previews within six hours, a verified partial set at twenty-four, and a final reconciled set at seventy-two.
One caution: do not deliver a file you have not opened. A JPEG that shows the wrong half of the frame, or a clip that plays for two seconds, is worse than nothing because it resets expectations and makes every later delivery suspect. The check is quick and the routine is in verifying recovered files.
Lab, Reshoot, or Nothing
Every card eventually lands in one of three buckets, and the deciding factor is arithmetic rather than sentiment.
| Situation | Route | Typical cost and time |
|---|---|---|
| Card mounts; files deleted or formatted over | Software scan from an image | Under an hour of your time |
| Card not detected, or visibly damaged | Data recovery laboratory | Hundreds to low four figures, days to weeks |
| Card crushed, submerged, or controller dead | Laboratory, with a real chance of nothing | Highest cost, longest wait |
| Footage can be reshot for less than the quote | Reshoot | Depends on logistics |
The comparison is not between the lab bill and zero. It is between the lab bill plus the waiting time and the cost of the alternative, which is usually a reshoot or a negotiated partial refund. Do not offer a reshoot before you know whether the data survives, because it closes the cheaper path and signals that you have given up. Offer it as a concrete option, with a date and a price, once you know it is the better route. The full comparison is in software versus a recovery lab.
Sometimes the honest outcome is that a card is unrecoverable. Deliver the parts that survived, itemize what was lost, explain the mechanism in one sentence, and put the remedy on the table before you are asked for it: reshoot at a reduced rate, a partial refund, or a credit against future work, whichever your contract supports. The criteria for that conclusion are in when data is unrecoverable.
If you are on Windows and the card is still detected by a computer, a purpose-built recovery tool such as XRecovery (xappset.com) can scan SD, microSD, and other memory cards for deleted or formatted photos and videos. As with any recovery tool, stop writing to the card first and always save recovered files to a different drive.
Frequently Asked Questions
Should I call the client as soon as I discover the loss?
Call when you have a diagnosis and a plan, which usually takes a couple of hours, not minutes. Never delay past a decision the client is about to make, such as booking a reshoot. And never conceal a loss to protect a deadline; that is the choice that turns a technical problem into a professional one.
Which card should I work on first when several are affected?
The one with the highest combination of client value and remaining intactness, which usually means the card nobody has touched since the loss. Image it before scanning. A card already formatted and refilled belongs last even if it matters most, because overwritten data cannot be recovered by any tool.
Can I deliver anything before the full recovery finishes?
Yes, and you should. Low-resolution proxies and the full-size JPEG previews embedded in RAW files often survive damage that destroys the main payload, so they can go to the client within hours. Follow them with verified full-resolution files as they land, and never send a file you have not opened.
What should I avoid promising a client?
Any specific recovery percentage, and any delivery date you cannot control. State what is intact, what is at risk, what you are doing, and when you will update them next. A figure you mention under pressure becomes the number the client measures you against, even when it was only ever a guess.