Blog Recruitment
What to record about a candidate so you can still find them in two years
Record three or four skills with their level, the real reason for the rejection and one sentence of context. With that, a search brings the person back two years later; without it you have a PDF in a folder and you pay for the advert again.
The problem is not storing. It is asking.
Almost every company keeps the CVs it receives. Very few can find them when they need to, and the difference is not where they are kept: it is that a folder of PDFs, an inbox or a spreadsheet of names cannot be asked anything. “Who do we have who speaks English and knows AutoCAD?” is not a question you can put to a pile of files.
The cost of not being able to ask does not show up because it is paid in instalments: every new vacancy is a new advert and a new screening round to reach, quite often, people you had already interviewed.
What you do need to record, and it is less than you think
Four things, recorded at the end of the interview, which is the only moment when you have them fresh and it costs two minutes:
- Three or four skills, with their level. Not fifteen. The ones that actually describe this person: “English C1”, “AutoCAD expert, 8 years”, “health and safety”. A level turns a tag into something you can filter on; a bare tag only says the topic rings a bell.
- The real reason for the rejection. “Did not fit the profile” and “we could not meet their number” look alike and are opposites: you never call the first one back, and you call the second the moment your band changes. Recording whether you ended it or they walked is the only way to know how many good ones you lose on money.
- One sentence of context. “Wanted three days remote”, “moving to Valencia in January”, “came recommended by Marta”. It is the first thing anyone forgets and the thing that most changes whether you call again.
- Where they came from. Without job-board integrations, the source is the only thing that answers where the people you actually hire come from — and that decides where the next advert budget goes.
What not to record
The temptation is to keep everything “just in case”, and it is a bad idea for two reasons. The first is legal: this is personal data about somebody who does not work for you, and data minimisation says you keep what the process needs and nothing more. Age, photo, marital status or nationality are not needed to decide whether somebody can weld.
The second is practical: a field nobody fills in properly is a field nobody can search on. A list of fifteen skills per person is half empty by the third interview; a list of three gets filled in every time.
And a loose opinion written in the heat of the moment — “did not like them” — is not information: two years later it explains nothing and, if the person asks for access to their data, you have to show it to them.
How you search afterwards
A talent pool you can question changes the order of the work: before advertising the vacancy, you search inside. And the searches that work have three properties worth demanding of any tool:
- The level has to be a minimum. Ask for English B2 and C1 speakers must come back too. It sounds obvious, and plenty of tools filter on an exact match, which rules out the best people.
- It has to return partial matches, ranked. Asking for five things and demanding all five returns nothing. What is useful is “matches 3 of 5” and which ones, and you decide.
- It has to tell you how many people there are before you search. Seeing a zero next to “German” saves you the search and tells you something more important: there is nobody there and it is time to advertise.
And the retention period, which has to be stated up front
Keeping a CV indefinitely “just in case” does not hold up. You set a reasonable period — a year is the most common for a selection process — you state it in the form itself before the person sends it, and you stick to it. What turns “we had their consent” into something you can show is storing the text exactly as it was worded on the day they accepted it, with the date.
We are not lawyers: if you have doubts about a specific case, check with whoever handles your data protection.
Frequently asked questions
How many skills should I record per person?
Three or four. That is what actually gets filled in at the end of an interview and it is enough to find somebody later. Fifteen fields per candidate are empty by the third interview, and an empty field is no use for searching.
What if the skill I need is not on the list?
You add it. The lists of languages, skills and qualifications are maintained by each company in its own words: a workshop puts “TIG welding” and an engineering firm puts “structural calculation”. An imposed dictionary with five hundred entries does not get used.
Is searching the words inside the CV not enough?
No. Searching for “English” inside a PDF returns the person who speaks it and the person who wrote it in their cover letter alike, and it cannot tell a B1 from a C2. Two minutes at the end of the interview are worth more than any full-text search.
Should I record why somebody was rejected?
Yes, and the real reason. It is the piece of data that most changes whether you call them back, and what separates a useful talent pool from a list of names. Record it at the time: six months later nobody remembers.
Can candidates be moved from one tool to another?
If the tool exports CSV or has an API, yes — and it is worth checking before you start putting data in, not on the day you want to leave. In Fichaloto candidates and their skills go in and out as CSV and through the API, and so do the skill lists.
All of this is built into Fichaloto. See how it works.
Keep reading
How to choose
Twelve questions worth asking before you sign
None of these twelve questions is technical. Every one of them is answered in a sentence, and you can ask for the answers in writing before signing anything.
Compliance
A clocking record that cannot be quietly rewritten
Most time-tracking tools store hours in a table anyone with database access can edit. That is fine until the day somebody asks you to prove a number.
Capacity
Planned, clocked, charged: three numbers that should match and usually do not
Most companies can tell you what they planned and what they invoiced. Very few can tell you, for a given week, whether those two things are related.