What a pay survey is
A pay survey (مسح الأجور), also called a salary survey, is a data collection tool: a single body gathers pay figures from several organisations for jobs described in a common way, then releases them in aggregate as percentiles, not as figures attributed to the organisations that supplied them.
The definition carries three conditions, and a misuse of a pay survey can be traced to dropping one of them. The collection is systematic, not impressionistic. The matching is done against a common job description, not against job titles. And the output is a distribution, not a single figure. Someone who gathers three figures from three acquaintances and calls the result a survey has dropped all three conditions at once.
What the output of a pay survey shows
The output is not an average. It is a distribution, examined at fixed points. Take one job for which twelve organisations reported pay:
| Point | What it means | Figure in this survey |
|---|---|---|
| 10th percentile | One tenth of those reported on are paid below this figure | 9,400 |
| 50th percentile (the median) | Half are paid below it and half above | 13,000 |
| 75th percentile | Three quarters are paid below it | 16,200 |
| 90th percentile | Nine tenths are paid below it | 21,000 |
These figures, and the others that follow, are arithmetic examples set up for explanation. They state no pay level for any job in the Saudi market, they describe no market, and they are not to be quoted as one.
An organisation that says “we pay at the 75th percentile” is saying that its pay is SAR 16,200 in this survey, with this sample, at this date. If three large organisations joined the sample in the next cycle, the figure would move without anything having changed in the labour market. The accurate statement carries its conditions: “at the 75th percentile of survey X for year Y, on a sample of twelve organisations”.
For the same reason, a percentile from one pay survey is not compared with a percentile from another. A percentile describes a position within a sample, and two different samples describe two different markets, even when they carry the same name.
Matching in a pay survey is half the tool, and the job title is no part of it
Survey figures are tied to the organisation through benchmark jobs (الوظائف المرجعية): jobs with defined content that exist in a similar form in other organisations, and so serve as points of reference. The match is made on the content of the work, not on the title.
The matching condition is stricter here than in any internal tool, because the other party is outside the organisation and nobody can check how it described its job. The general rule of benchmarking (المقارنة المعيارية) applies: a comparison holds only when both sides measure the same thing in the same way.
An error in matching weighs more in a pay survey than in other tools, because it does not show. The figure comes out of the survey ordered and calculated in percentiles, and it looks precise whether the job was matched to its true counterpart or to a job that shares its title but not its work. Arithmetic precision does not correct a faulty match.
The sample of a pay survey is defined before its figures are examined
“The market” in a pay survey is not one thing. It is whatever the sample defines. Three constraints are set before any figure is examined: activity, organisation size and location. Changing any one of them moves the median without the market moving. Location carries weight for the reason behind geographic pay differentials: the range set for one job can vary from one city to another.
Take the same job in the same survey, examined under three definitions of the sample:
| Sample definition | Organisations | Median |
|---|---|---|
| All reporting organisations | 12 | 13,000 |
| Large organisations only | 5 | 16,800 |
| Organisations in the same activity only | 4 | 11,500 |
The gap between the highest reading and the lowest is SAR 5,300, in a single survey, for a single job, on a single day. The right decision does not start from the question “what is the figure” but from the question “whom do we compare ourselves with, and why”. Where an organisation competes for candidates with others in the same activity, the third row is the relevant sample, even though four organisations are too few to carry a decision by themselves, a point taken up below. Where its employees leave for large employers, whatever their activity, the second row is. Taking the first row because it is “the broadest” means looking at the median of a market in which the organisation does not compete.
A pay survey reports pay that is paid, not pay that is advertised
Another source can be confused with a pay survey: it resembles one in its output but differs in its origin. That source is a compilation of the figures published in job advertisements. The difference between the two is not one of precision; it lies in what each describes.
- A pay survey. It describes what is actually paid to people in post, including those who have spent years in their roles.
- Job advertisements. They describe what organisations offer to new entrants, which is one segment of the distribution and not the whole of it, and they may include offers that were never accepted.
So the advertised figure can rise above the survey figure in a market where competition to hire is intense. That rise is real, but it is a rise in the price of entry, not in the price of staying. An organisation that corrects its pay table on advertised figures alone corrects it on the upper end of a distribution it has not seen in full. Where entry pay is raised and the pay of serving employees is not, the gap between the two can narrow, which is the ground of pay compression.
The effective date of a pay survey, and ageing the data
Every pay survey has an effective date: the day on which its data describe pay. It is neither the date on which the report is published nor the date on which it is consulted. The gap between the effective date and the date of the decision is handled by ageing the data (تقادم البيانات): the figure is multiplied by a movement factor for the intervening period.
Take the median above, SAR 13,000, with an effective date of 1 January, a decision taken on 1 July, and an assumed movement factor of 4% a year:
- The intervening period is six months, or half a year.
- Movement for the period = 4% × 0.5 = 2%.
- Aged figure = 13,000 × 1.02 = SAR 13,260.
One error at this step is double counting. The survey provider has already aged the figures to a future date in its report, and the organisation then ages them again from the original effective date. The figure comes out at 13,000 × 1.02 × 1.02 = SAR 13,525.20, which is SAR 265.20 higher with nothing to account for the difference. The safeguard is one line written beside every figure: the date to which it has been aged.
The data sent to a pay survey are personal data
What an organisation sends to a survey provider are not bare figures. They are the pay of identified individuals, held by their employer. The Personal Data Protection Law (نظام حماية البيانات الشخصية) attaches obligations to the processing of personal data. Among them, the purpose of collection is specified, and the data are not processed for another purpose except in the situations the law lists, consent being one of them. The content of the data is limited to the minimum needed to achieve that purpose. Sufficient steps are taken to verify that the data are accurate, complete, current and relevant to the purpose before they are processed. And the data are destroyed once they are no longer needed for the purpose.
The practical consequence is simple: a pay survey does not need names. What the survey requires is the job description and the pay that goes with it; an employee’s name and staff number serve no purpose in an aggregated output. An organisation that sends a full payroll extract because it is easier than preparing a reduced file sends more than the purpose requires. Removing names does not on its own take a file outside the law, because the law’s definition of personal data covers data that identify a person indirectly as well as directly.
Which of the situations listed in the Personal Data Protection Law, if any, covers sending these data to a third party, and what procedure that requires, we do not settle. In the sources we reviewed, we found nothing that decides the point, and it is a matter for the law itself and its executive regulations.
What a pay survey is not
- The comparable wage (أجر المثل). It is a rule for filling a gap under Article 95 of the Saudi Labor Law (نظام العمل), called upon where the wage is stated neither in the contract nor in the work regulation (لائحة تنظيم العمل). It applies when a stated wage is absent, whereas a pay survey is a management tool used when a wage is being set or reviewed. Presenting the output of a pay survey as “the comparable wage” puts a commercial figure where a statutory rule belongs.
- Benchmarking. It is broader and more general than a pay survey: it compares the performance of an organisation, or of one of its processes, with an external or internal reference, and its aim is to identify a gap and its source. A pay survey is one of its inputs when the subject of the comparison is pay.
- Pay equity (عدالة الأجور). It looks inside the organisation, at explaining the differences in pay between its employees. A pay survey looks outside it, at its position against others. An organisation at the 90th percentile may carry an evident imbalance inside it, and the reverse is equally possible: neither tool replaces the other, and neither reveals what the other reveals.
What makes a pay survey sound to build on
- The number of organisations reporting each job, published beside its figure. It matters because a percentile calculated on four organisations is too thin to carry a decision.
- The job description used for matching, published in full rather than reduced to a title. It lets the reader judge the match for themselves.
- A defined pay base. It states whether the figure is basic pay or includes allowances and variable pay. Comparing one base with another is an error that can be hard to detect once it has happened.
- The effective date and the ageing status, as set out above.
Of these four, the pay base carries the heaviest consequence. That is not because the amount involved is larger, but because its error runs in one direction and falls on every line. An organisation that sent only the basic wage (الأجر الأساسي) for its employees and then took the survey output to be the actual wage (الأجر الفعلي) believes it is below the market when it is in line with it. If basic pay were SAR 9,000 and allowances SAR 2,600, the gap between the two readings would be SAR 2,600 on every line of the table, always in the same direction: no line offsets another, and no job corrects the error made on its neighbour.
That amount is in itself smaller than the distance between percentiles in the first table: SAR 3,600 between the 10th percentile and the median, SAR 3,200 between the median and the 75th percentile, and SAR 4,800 between the 75th and the 90th. So the argument does not rest on size. The difference is that choosing a percentile is one decision, taken once, written down and revisited in the next cycle, and whoever took it knows the direction and size of its effect. An error in the pay base is a silent bias that runs through every job measured and every comparison built on them, and it announces itself on no line. That is why the pay base is settled first, before the organisation is asked at which percentile it wants to stand.
Before buying a pay survey
In the sources we reviewed, we found nothing establishing that a government body publishes an official pay survey by occupation. Equally, we do not state that no such survey exists.
The limits on exchanging pay data between organisations under competition rules (أنظمة المنافسة) are a separate matter. In the sources we reviewed, we found no text that addresses them, so we state nothing on them in either direction.
A pay survey informs a decision about pay; it does not take one. Its findings can change the size of the amount a salary review shares out, but they do not divide that amount. We recommend no provider. An organisation buying a survey can ask three questions before it asks about price: how many organisations reported on the jobs that matter to it, which pay base the figures carry, and to which date they have been aged.
This is an explanation of the concept and of the statutory provisions cited, not legal advice.
A standalone Saudi HR system
One employee file holding the contract, the documents and their expiry dates, the attendance record, leave, salary and end-of-service entitlements. End-of-service, overtime and leave-balance calculations are built into the system.
A standalone system on its own subscription. The connection to Qoyod Accounting is now available.