People analytics for small businesses: the data you already have
By Deibson Silva · Reviewed on
People analytics for small businesses starts with the data the company already keeps: joiners and leavers, vacancies, performance reviews, training and surveys. With it you can answer four or five business questions, always with a stated purpose, with groups of at least five people and without borrowing data that was collected for something else.
What is people analytics in a small company?
It is using data about your workforce to make one specific decision better. In a large company that sounds like statistical models. In a business of forty people, nearly all the work is counting properly and asking the right question.
The small company's advantage is that it knows its people and can check any figure by talking to them. The risk is the opposite: with so few cases, a single number misleads easily.
What data does a small business already have?
More than it thinks, spread between the accountant, the inbox and several spreadsheets. This table lists the usual items, where they tend to sit and what each one is good for.
| Data | Where it usually sits | Question it answers | Watch out for |
|---|---|---|---|
| Start and leaving dates, and the reason for leaving | Payroll package or accountant | Who leaves, when and from which area | With few leavers, count cases instead of percentages |
| Vacancies: opening, closing and each candidate's source | Email and spreadsheets | How long it takes to fill a role and which channel works | Keep only what you need about people you did not hire |
| Performance reviews | Each manager's documents | Which teams have gone more than a year without a review | Compare only what was rated against the same criteria |
| Training completed | Invoices and certificates | Which teams train and which do not | Do not cross it with personal ratings without saying so |
| Pay register | Accountant | Pay differences by role and by sex | In Spain the law already says how to break it down |
| Engagement surveys | Online forms | What worries each area | Without a minimum number of responses per group there is no anonymity |
Which data can you not use, even if it is to hand?
The clearest case in Spain is the daily record of working time. Every company keeps one because Article 34.9 of the Workers' Statute requires it, and it is tempting to use it to measure who delivers most. The Spanish data protection authority's guide on employment rules that out: data from the record cannot be used for purposes other than checking working time.
It is the general rule in Article 5 of the GDPR applied to one case. Data is collected for specified purposes and is not then processed in a way incompatible with them. It must also be adequate, relevant and limited to what is necessary.
The special categories in Article 9 stay out as well, such as trade union membership or the medical reasons for an absence. People analytics does not need them and they should not go into the sheet.
The pay register is a different case. Article 28.2 of the Workers' Statute obliges Spanish companies to keep one with the average values of salaries, supplements and non-salary payments, broken down by sex and by group, category or role. The law already asks for that analysis, and it is worth doing well.
Which questions can that data answer?
Five questions are enough for a year's work. Each ends in a decision.
How long do we take to fill a role, and at which stage does it get stuck? The dates of each vacancy show whether the problem lies in finding candidates or in deciding between them.
Who leaves in their first year, and from which area? If early departures cluster in one team, look again at how those people were selected and how they were welcomed.
Which roles have nobody ready to step in? All it takes is a list of roles and, next to each, the name of who could take it on. The empty boxes are the year's development plan.
Which teams have gone more than a year without a performance review? The answer usually points to managers who need support to give one.
What did each area say in the last survey, and what did we do afterwards? If nothing was done, the next one will get fewer answers.
How do you work with small numbers?
By counting cases. Saying that turnover in a team of six is 33% sounds alarming; saying that two people left, one of them because they moved city, explains what happened.
Write the definitions before you count. What a voluntary departure is, when a vacancy counts as filled, what counts as a performance review. Use the same definition every quarter, because changing it mid-year makes comparison impossible.
Do not show results for groups of fewer than five people. In a group of three, anyone can work out who said what. If a team is very small, combine it with another.
And do not use the data to predict what one particular person will do. With this number of cases there is no basis for it, and a direct conversation with them tells you more.
What do you have to tell staff?
Which data is used, why and who sees it. Article 13 of the GDPR requires you to tell each person who the controller is and the purposes their data is intended for. A one-page note is enough, if it is clear.
In Spain, if the company has a works council and uses algorithms that influence employment decisions, Article 64.4.d of the Workers' Statute entitles the council to know their parameters, rules and instructions.
When the analysis includes performance, personality or behaviour, the Spanish authority's list places it under its first criterion, and expects an impact assessment in most cases where two or more criteria are met. Check this before you start. These are Spanish and EU rules; other jurisdictions have their own.
When does the spreadsheet stop being enough?
When more than one person edits it and nobody knows which version is the good one. Every copy is also a copy of personal data that someone has to control.
It also stops being enough when decisions start to combine data. To know who can take on a role you need their performance, their potential and what they still have to learn, and that does not fit on one tab.
At that point a platform makes sense: one that brings the data together, applies the minimum number of responses per group by itself and records who changed what.
How do you start without buying anything?
Pick one of the five questions, the one that worries senior management most at the moment. Gather the data in a single sheet, with its definition written at the top. Name a person to maintain it and go through it with senior management once a quarter.
When that question has an answer and has helped to decide something, add the second. The order matters less than the habit of looking at the data before deciding.
Sources
- Regulation (EU) 2016/679 (GDPR), Articles 5, 9 and 13 (EUR-Lex)
- AEPD: data protection in employment relationships, section 8, working-time records (in Spanish)
- Spanish Workers' Statute, Articles 28.2, 34.9 and 64.4.d (BOE-A-2015-11430, in Spanish)
- AEPD: list of types of processing that require a data protection impact assessment (Article 35(4), in Spanish)
Frequently asked questions
- Do you need a data analyst?
- No. A small business needs one tidy person who keeps a sheet with clear definitions, and a management team that looks at it before deciding.
- How many people do you need for it to make sense?
- It makes sense from the moment a decision about staff repeats. With twenty people you can already measure how long it takes to fill a role or which teams have had no performance review.
- Can I use clocking-in data?
- Not for people analytics in Spain. The data protection authority's guide on employment says that data from the working-time record cannot be used for purposes other than checking working time.
- How often should you look at the data?
- Once a quarter is enough for a small business. Less often and it gets forgotten; more often and the numbers barely move.
- How is it different from an HR dashboard?
- A dashboard shows indicators. People analytics starts from a business question and uses only the data that helps to answer it.