Session 1: Data Discipline

Data entry is the transfer of information from one place to another — paper to spreadsheet, email to a CRM, one system to a better one, an image to text — where the value lies in the transfer being correct. It is paid work in Nigeria in real volume: hospitals digitising patient records, schools entering results, logistics firms capturing waybills, e-commerce sellers loading products, accounting firms keying invoices, research projects transcribing survey responses, and remote virtual-assistant work for foreign clients who pay in dollars.

What clients pay for is not keystrokes. It is the absence of errors. A retail inventory sheet with 200 wrong prices costs a business real money; a payroll sheet with one misplaced digit is a crisis; a research dataset with inconsistent entries is unusable and must be redone from source. This is why the professional standard is measured in accuracy percentage rather than speed, and why a slower, careful worker keeps clients while a fast, careless one loses them after the first audit. Understand that trade and you understand the whole job.

Every piece of data has a type, and the type determines what can be done with it. Text is anything read as characters — names, addresses, product descriptions. Numbers can be calculated on. Dates can be compared and subtracted. Currency is a number with a money format. Boolean is a yes/no or true/false flag. Identifiers — phone numbers, account numbers, ID numbers, invoice references — look numeric but are text, because you never add two phone numbers together and a leading zero is meaningful.

Getting the type wrong is the single most common data entry failure and it fails silently. Type a phone number as a number and Excel drops the leading zero and may show it in scientific notation; type a date as text and no date calculation will ever work on it; type an amount with a naira symbol by hand and it becomes text that SUM ignores. The visible symptom is a total that does not add up or a sort that produces nonsense, and by then the error is scattered across hundreds of rows. Learn to enter each type correctly at the point of entry and you eliminate the entire category.

The vocabulary matters because it is how you describe problems precisely. A field is one category of information — surname, phone number, amount. A record (or row) is one complete entry across all fields — one customer, one transaction. A table is a set of records sharing the same fields. A form is the interface for entering one record at a time, and it is safer than a raw table because it shows one record, labels each field, and can validate as you type.

The rule that follows is one fact per field. Never enter 'Adebayo Okafor — 0803 456 7890 — Lagos' into a single cell. Split it into Name, Phone and City. Combined fields cannot be sorted, filtered, searched, deduplicated or reported on, and every client request that requires any of those becomes a manual re-typing job. If you receive data already combined, splitting it is often the first task you are being paid for.

Accuracy is the proportion of fields entered correctly. Speed is throughput, usually records per hour or keystrokes per hour. They trade against each other, and the professional order is accuracy first, speed second — because speed is easy to gain with practice and accuracy is a discipline you either have or do not.

Realistic thresholds in Nigerian and remote work: general business data should be at 98% field accuracy minimum, with 99.5% or better being what makes a client keep you. Financial fields — amounts, account numbers, invoice references — should be at 100%, because a single error moves money. Speed expectations for general entry sit around 35–45 words per minute equivalent, or roughly 150–250 simple records per hour depending on field count. Note what those numbers imply: at 98% accuracy across 1,000 fields you have made twenty errors, which is why verification is not optional and is the subject of session three.

Errors are not avoided by concentration. Concentration is a finite resource and it fails on the four-hundredth row of a long session regardless of how careful you were. Errors are avoided by process: entering in a fixed order so you never skip a field, verifying in batches rather than at the end when the source is far away, checking totals against the source document, and reviewing at natural breakpoints rather than after hours of continuous entry.

Three habits do most of the work. First, batch and verify: enter twenty or fifty records, then check them against the source before continuing — errors found immediately are cheap, errors found after five hundred rows require re-reading everything. Second, spot-check by column: after a batch, scan one field down the whole column; inconsistencies in format, spelling or type jump out vertically in a way they never do horizontally. Third, stop when you are tired. Error rates climb steeply with fatigue, and the last hour of a long session produces most of the mistakes. Take a real break every forty-five minutes.