Session 4: Files, Naming & Cloud Backup
The layer that protects everything else: a folder structure you can navigate, filenames that survive being separated from their folder, cloud storage used properly, a backup strategy that assumes devices die, and how to recover a lost file.
Learning objectives
By the end of this session you will be able to do each of these without prompting.
- Build a folder structure shallow enough to actually use
- Apply a naming convention that works when a file leaves its folder
- Use cloud storage as a working system rather than a dumping ground
- Implement the three-two-one backup principle on a real budget
- Recover a deleted or overwritten file, and know when recovery is not possible
The taught content
The problem is not storage, it is finding
Nobody runs out of space; people run out of ability to locate what they saved. Storage is cheap and effectively unlimited, so the failure mode of personal file management is not loss of capacity — it is a document you know exists and cannot find, or three versions of it with no way to tell which is current. That is the problem this session solves, and it is why naming and structure matter more than any amount of storage.
The cause is almost always the same: files were saved at the moment of saving rather than with any system, into whatever location was convenient, with whatever name came to mind. 'Document1', 'final', 'new version', 'invoice' — none of which mean anything six weeks later. A filename is a message you write to your future self, and most people write nothing.
There is also the deeper risk that this course has been circling: a file on one device is one hardware failure away from gone. Laptops die, phones are stolen, and disks fail without warning. A system with no backup is not a system, it is a countdown — which is why the second half of this session is about redundancy rather than organisation.
Folder structures: shallow beats deep, every time
The instinct is to build a detailed hierarchy, and the result is a tree nobody can navigate. Three or four top-level folders, with at most two levels beneath them, is the structure that actually gets used — because you can hold it in your head, and because filing a document requires one decision rather than five.
Organise by category, then by time, rather than trying to build a perfect taxonomy. Work, Personal, Finance, Documents at the top level; then by year or by client within them. Deeply nested structures fail for a specific reason: every additional level is another decision at the moment of filing, and when filing is effortful, people dump files on the Desktop instead — which is how a careful structure ends up empty while the Desktop holds everything.
The supporting habit is a single inbox for the undecided. A folder called To Sort, checked weekly, catches the files you cannot immediately categorise. Without it, an unclear file goes to the Desktop and stays there; with it, filing becomes a small weekly task rather than a source of friction every time you save something.
Naming conventions: the filename must survive leaving its folder
The test of a good filename is this: if the file were emailed to you with no folder context, would you know what it is? 'invoice.pdf' fails. '2026-03-14 Adeyemi Interiors — Invoice 0042.pdf' passes. Files get attached, downloaded, shared and copied out of their folders constantly, and the moment that happens, the folder structure you relied on is gone and the name is all that remains.
The convention that works has three parts. Date first, in year-month-day order — 2026-03-14 — because it sorts chronologically in every file manager without any configuration, which no other date format does reliably. Then who or what it concerns, in plain words. Then what it is. That order puts the most useful sorting key first and reads naturally left to right.
Two rules complete it. Never use the words final, new, or a version number as the distinguishing part — 'final_v2_REAL' tells you nothing except that you were unsure, and it guarantees ambiguity. And be consistent rather than clever: a mediocre convention applied to every file beats an elegant one applied to half of them, because search only works predictably when naming is uniform.
Cloud storage: a working system, not a dumping ground
Cloud storage does three distinct jobs and it is worth separating them. It syncs your files across devices, so the phone and the laptop see the same things. It provides access from anywhere, which is what makes a dead laptop survivable in the short term. And it keeps version history, so an overwritten file can be restored. Those are three different protections and none of them is a substitute for a real backup.
The habit that makes cloud storage work is saving to it by default rather than to the local disk. Most people save locally and upload occasionally, which means the cloud copy is always stale and the local copy is always at risk. Invert it: the cloud is where the file lives, and the local copy is a cache. Once that is your default, the sync happens without thought and the risk profile changes completely.
Then the caution that matters. Sync is not backup. If you delete a file, or overwrite it with something wrong, or ransomware encrypts it, the change syncs to the cloud just as faithfully as an intentional edit. Cloud storage protects against device failure and nothing else. Treating sync as backup is the single most common and most expensive misunderstanding in personal file management.
Backup: the three-two-one principle, and recovering what is lost
Three copies of your data, on two different types of medium, with one kept offsite. That is the standard, and it exists because each part covers a failure the others do not. Three copies because any single copy can fail unnoticed. Two types of medium because a disk fault and a cloud account problem are different risks. One offsite because theft, fire and flood take the building and everything in it.
On a realistic Nigerian budget this is achievable without expense. Copy one is your working files in cloud storage. Copy two is a scheduled export to an external drive or a second cloud service. Copy three, offsite, is a second cloud provider or a drive kept somewhere other than where the computer lives. The critical part that most people skip is the schedule — a backup you must remember to run will be missing on the day you need it, so automate it or attach it to something you already do weekly.
Then recovery, which is the part that determines whether a backup is real. Cloud version history restores an overwritten or deleted file within the retention window. Cloud trash holds deleted items for a limited period before permanent removal. A drive backup restores anything from the moment it was taken. And the honest limits: a file deleted from everywhere, past the retention window, with no backup, is gone — and knowing that in advance is what motivates the schedule. Test a restoration once, deliberately, while nothing is at stake, because a backup you have never restored is a hope rather than a plan.
Instructor demonstration
We organise a real set of files, apply a naming convention, invert the save-to-cloud default, build a three-two-one backup on a real budget, and recover a deliberately destroyed file.
- 01
Survey what you actually have before organising anything
Count the files on your Desktop and in Downloads, and look at how many are named 'new', 'final', 'document' or a date you cannot interpret. This is the baseline, and it explains why finding things takes so long.
- 02
Create three or four top-level folders
Work, Personal, Finance, Documents. Resist adding more — the structure that gets used is the one you can hold in your head, and every extra category is another decision at the moment of filing.
- 03
Add at most two levels beneath them
By year or by client within each. Deeper than that and filing becomes effortful, which is precisely when people start dumping files on the Desktop instead.
- 04
Create a To Sort folder as the inbox for the undecided
For anything you cannot categorise immediately. Without it an unclear file goes to the Desktop and stays there; with it, filing becomes one small weekly task rather than friction every time you save.
- 05
Move your Desktop and Downloads contents into the structure
Do it in one sitting, putting anything unclear into To Sort rather than agonising. A cleared Desktop is the visible proof that the system works, and it takes less time than deciding each file individually.
- 06
Rename ten files to the convention
Date first as 2026-03-14, then who or what it concerns, then what it is. Note how much more searchable they become, and how they now sort chronologically without any configuration.
- 07
Apply the email test to each name
Ask whether the filename alone would tell you what the file is if it arrived with no folder context. Files leave their folders constantly, and when they do the name is all that remains.
- 08
Remove every 'final', 'new' and version number
Replace them with content words. 'final_v2_REAL' tells you nothing except that you were unsure, and it guarantees you will not know which copy to use six weeks later.
- 09
Move your working files into cloud storage
Not a copy — the files themselves. The cloud becomes where they live and the local disk becomes a cache, which inverts the usual arrangement where the cloud copy is always stale.
- 10
Confirm sync is actually working on both devices
Create a file on the laptop and find it on the phone. Create one on the phone and find it on the laptop. Assuming sync works is how people discover it does not, at the worst moment.
- 11
Demonstrate why sync is not backup
Delete a file and watch it disappear from the cloud too. This is the moment the distinction becomes concrete: sync faithfully replicates mistakes, ransomware and accidental deletion.
- 12
Set up backup copy two
An external drive, or a second cloud service, holding a scheduled export of your working files. Two different types of medium, because a disk fault and an account problem are different risks.
- 13
Set up backup copy three, offsite
A second cloud provider, or a drive kept somewhere other than where the computer lives. Theft, fire and flood take the building and everything in it, so offsite is not optional.
- 14
Attach the backup to a schedule you will actually keep
Automate it, or tie it to something you already do weekly and put that in your calendar. A backup you must remember to run will be missing on the day you need it.
- 15
Recover a file from cloud version history
Overwrite a document badly, then restore the earlier version. Doing this deliberately while calm is what makes it a skill you have rather than a feature you have heard about.
- 16
Recover a file from cloud trash
Delete something and restore it from trash, noting the retention period. After that window, deletion becomes permanent — which is a fact worth knowing before you need it.
- 17
Restore a file from the drive backup
Prove the second copy works by restoring from it. A backup you have never restored is a hope rather than a plan, and this test takes two minutes.
- 18
Establish the honest limit
Name the situation where recovery is impossible: deleted everywhere, past the retention window, no backup. Knowing that precisely is what motivates keeping the schedule.
- 19
Write your conventions down
The folder list, the naming pattern, the backup schedule and the recovery steps. A system that exists only in your head stops existing the day you are busy or the day someone else needs to help you.
- 20
Confirm the complete system works end to end
Save a new file with the convention into the structure, confirm it syncs, confirm the backup captures it, and confirm you can find it by search a week later. That is the deliverable: a system, not a tidy folder.
Guided practice
Build the system that protects everything else
Organise your real files, apply a naming convention that survives a file leaving its folder, build a three-two-one backup on a real budget, and prove you can recover.
- 01Survey your Desktop and Downloads, counting files with uninformative names as your baseline.
- 02Create three or four top-level folders with at most two levels beneath them.
- 03Create a To Sort folder for files you cannot categorise immediately.
- 04Move everything from Desktop and Downloads into the structure in one sitting, using To Sort rather than agonising.
- 05Rename ten files to the convention: date as 2026-03-14, then who or what, then what it is.
- 06Apply the email test to each name — would the filename alone tell you what it is?
- 07Remove every final, new and version number, replacing them with content words.
- 08Move your working files into cloud storage so the cloud is where they live.
- 09Confirm sync works in both directions between laptop and phone.
- 10Delete a file and watch it vanish from the cloud, proving sync is not backup.
- 11Set up backup copy two on an external drive or a second cloud service.
- 12Set up backup copy three offsite, on a second provider or a drive kept elsewhere.
- 13Attach the backup to automation or to a weekly habit recorded in your calendar.
- 14Restore an overwritten file from cloud version history.
- 15Restore a deleted file from cloud trash and note the retention period.
- 16Restore a file from your drive backup to prove the second copy works.
- 17Write down the folder list, naming pattern, backup schedule and recovery steps.
- 18Save a new file with the convention and confirm it syncs, is backed up, and is findable by search.
The standard we hold you to
A baseline count of uninformatively named files; three or four top-level folders with at most two levels and a To Sort inbox; Desktop and Downloads emptied into the structure; ten files renamed with date-first convention, each passing the email test, with no final/new/version wording; working files living in cloud storage with sync verified in both directions; the sync-is-not-backup distinction demonstrated by deleting a file; **three-two-one backup implemented with copy two and an offsite copy three, attached to a schedule that will actually be kept**; recovery demonstrated from cloud version history, cloud trash and the drive backup; the honest limit of recovery stated; conventions written down; and a new file confirmed to sync, be backed up and be findable by search.
Common mistakes and how to fix them
Building a deep folder hierarchy nobody can navigate
Fix: Three or four top-level folders with at most two levels beneath. Every additional level is another decision at the moment of filing, and effortful filing sends files to the Desktop instead.
Naming files for the folder they are in
Fix: Files get attached, downloaded and copied out constantly, and then only the name remains. Apply the email test: would the filename alone tell you what it is?
Using final, new, or version numbers as the distinguishing part
Fix: 'final_v2_REAL' records your uncertainty, not the content. Use date plus who or what plus what it is, and let version history handle drafts.
Saving locally and uploading occasionally
Fix: That leaves the cloud copy stale and the local copy at risk. Invert it: the cloud is where the file lives and the local disk is a cache.
Treating cloud sync as backup
Fix: Sync faithfully replicates deletion, overwriting and ransomware. It protects against device failure and nothing else. The most common and most expensive misunderstanding in personal file management.
Having one copy of anything important
Fix: Three copies, on two types of medium, with one offsite. Each part covers a failure the others do not, and a single copy can fail unnoticed.
A backup with no schedule
Fix: A backup you must remember to run will be missing on the day you need it. Automate it or tie it to an existing weekly habit and put it in the calendar.
Never testing a restoration
Fix: A backup you have never restored is a hope rather than a plan. Restore something deliberately while nothing is at stake, from each copy, and note the retention windows.
Expert notes
The habits that separate someone who can do this from someone who does it well.
- Name files so the name survives leaving the folder. Files are attached, downloaded and copied out constantly, and the folder structure you relied on is gone at that moment — the filename is a message to your future self, and most people write nothing.
- Use date-first naming in year-month-day order. It is the only date format that sorts chronologically in every file manager without configuration, which is a small detail with a permanent daily payoff.
- Sync is not backup, and this is worth internalising before it costs you something. Cloud storage faithfully replicates an accidental deletion or an overwrite, so it protects against device failure and nothing else. You need a separate copy.
- Test a restoration deliberately while nothing is at stake. Almost everyone who has done it found something they had assumed about their backup; almost everyone who has not done it finds out during the incident. Two minutes now is cheap.
Key terms
- Shallow structure
- A folder hierarchy of three or four top-level folders with at most two levels beneath. Deep trees are not used because filing becomes effortful.
- To Sort folder
- An inbox for files that cannot be categorised immediately, reviewed weekly. Without it, unclear files accumulate on the Desktop permanently.
- Naming convention
- A consistent filename pattern — date as YYYY-MM-DD, then who or what it concerns, then what it is. Its test is whether the name works with no folder context.
- The email test
- Asking whether a filename alone would identify the file if it arrived as an attachment. Files leave their folders constantly, so this is the real standard.
- Sync versus backup
- Sync replicates changes both ways including deletions and overwrites; backup keeps an independent copy. Sync protects against device failure only.
- Three-two-one principle
- Three copies of your data, on two types of medium, with one kept offsite. Each element covers a failure mode the others do not.
- Version history retention
- The period during which earlier versions and deleted files can be recovered. After it, removal becomes permanent.
- Test restoration
- Deliberately restoring a file from each backup copy while nothing is at stake. The only real evidence that a backup is usable.
Homework before the next session
Reorganise your real files
Three or four top-level folders with at most two levels, a To Sort inbox, and Desktop and Downloads emptied into the structure in one sitting. Record how many files had uninformative names before you started.
Rename twenty files to your convention
Date-first as YYYY-MM-DD, then who or what, then what it is. Apply the email test to each and remove every final, new and version number. Note how much faster search becomes.
Implement three-two-one backup
Working files in cloud storage, copy two on a drive or second service, copy three offsite — attached to a schedule you will actually keep, recorded in your calendar.
Prove you can recover
Restore a file from cloud version history, one from cloud trash, and one from your drive backup. Note each retention window and write down the situation in which recovery would be impossible.
Assessment rubric
How this session is marked. The certificate for Digital Productivity is awarded on the deliverable, not on attendance.
| Criterion | Passing | Excellent |
|---|---|---|
| Structure | Folders were created. | Three or four top-level folders with at most two levels, a To Sort inbox, and Desktop and Downloads genuinely emptied rather than partially tidied. |
| Naming discipline | Some files were renamed. | Date-first YYYY-MM-DD convention applied consistently, each name passing the email test, with no final/new/version wording anywhere. |
| Cloud usage | Files are in the cloud. | Cloud is where files live rather than an occasional upload, sync verified in both directions, and the sync-is-not-backup distinction demonstrated rather than merely stated. |
| Backup completeness | A backup exists. | Three-two-one implemented with an offsite copy, attached to a schedule that will actually be kept, and each copy verified by restoring from it. |
| Recovery capability | Knows where deleted files go. | Has restored from version history, trash and the drive backup, can state each retention window, and knows precisely the situation in which recovery is impossible. |
Session questions
Is cloud storage enough on its own?+
No. Sync faithfully replicates deletion, overwriting and ransomware, so it protects against device failure and nothing else. You need an independent second copy and an offsite third — that is what three-two-one means.
How much does a proper backup cost?+
Very little. One external drive plus a second cloud account covers three-two-one for a personal system. The expensive part is not the equipment; it is the schedule, which costs nothing but requires a habit.
I have thousands of badly named files. Do I rename them all?+
No. Rename files as you touch them, and start with the twenty you use most often. A convention applied going forward is worth more than a weekend of retroactive renaming you will abandon halfway through.
Can I always recover a deleted file?+
No, and the limits are worth knowing in advance. Cloud trash and version history have retention windows; after those, and with no separate backup, the file is genuinely gone. That is the argument for the schedule.
Why date-first naming rather than putting the date at the end?+
Because YYYY-MM-DD sorts chronologically in every file manager without configuration, and no other date format does so reliably. It is a small choice with a permanent payoff every time you look at a folder.
This session is part of
Digital Productivity
2 weeks · 4 sessions · ₦30,000 · you leave with a working personal system