Notes for owners · Branding and communication
How to measure whether your employer branding is working
Count applicants by source, offer acceptance and early exits from your own records, with a simple register layout and a rule for when the numbers are too small to read.
The GullySales team · Updated 17 Oct 2026 · 5 min read
On this page
- One row per person, filled the day something happens
- Each stage points at a different suspect
- When are the numbers too small to trust?
- For example: an embroidery unit in Surat
- How do you ask the source question?
- When does a reason become a decision?
- What should stay out of the register
- Back-fill your last ten hires
The question behind "is our employer branding working" is smaller and more useful: where did the people we wanted go? You can answer it from records you already hold. Note where each applicant heard of the job, whether they came when invited, whether they accepted the offer, and whether they stayed through the first appraisal. One person keeps a plain register. You need no software and no benchmark, only the same columns filled in for every candidate and the discipline to compare yourself with your own earlier record.
One row per person, filled the day something happens
| Name | Role | Heard of the job from (their words) | Applied | Shortlisted | Came | Offered | Joined | Left (date, reason in a line) |
|---|---|---|---|---|---|---|---|---|
| Example | Machine operator | "My cousin works there" | date | Yes | Yes | Yes | Yes | After first appraisal; shift longer than described |
Two columns do most of the work: where the person heard of the job, and why they left. Write both in the person's own words. After twenty rows you may see a pattern that no score would have shown.
Pick one definition of "early exit" and keep it fixed, for example leaving before the first appraisal. If you change it halfway, you cannot compare the two halves.
Each stage points at a different suspect
Read the register stage by stage, never as one grand figure. Here is what each break can mean.
| Where people drop | First place to look |
|---|---|
| Many apply, few shortlisted | The post reaches the wrong people, or describes the role loosely |
| Shortlisted, few come | Slow reply or a vague visit arrangement; see the ten moments |
| Offered, few accept | The offer differs from the advert, or pay is below the town's |
| Joined, leave early | Onboarding, supervision, or a promise the job did not keep |
Look at the source column last. If current employees and friends bring the people who stay, your name is travelling by word of mouth, which is the point of employer branding. If paid posts bring the volume and friends bring the people who stay, spend less on volume.
When are the numbers too small to trust?
Below about ten hires in one role family over the period, treat the numbers as a list of stories, not a trend. One person leaving moves everything. Keep the register anyway and read the reasons, which are useful at any size.
State the base whenever you say a rate. "Three of seven accepted" tells you more than a percentage. Compare a role only with its own earlier period, in the same town, hired the same way. A drop in applications after a festival or a jump because a rival closed a plant says nothing about your branding. And never set your figures against a published number from another sector or city.
For example: an embroidery unit in Surat
Take an invented unit that hires machine operators. It puts posters near two bus stops and posts in two local groups. After three months the register has twenty-one rows. Fourteen applicants came from the groups, five from friends of current operators, two from the posters. Of the seven who were offered the job, four joined. Of those four, three were brought by operators and are still there after the first appraisal. The three who declined all gave the same reason in the "reason" column: the poster said a nine-hour shift and the owner confirmed ten on the phone.
The owner's first thought was to spend more on the groups, because they brought the most people. The register points elsewhere. The poster is corrected to the true shift, operators are asked to bring a friend, and the groups become a way to get the first call. Next quarter the owner reads the new rows against these. The fix has a cost: the ten-hour shift is now printed, and the owner expects fewer applicants.
The counts are small, which is why the owner reads reasons and quotes "four of seven", not a percentage. The mismatch between poster and phone call would have shown up even at five rows.
How do you ask the source question?
Ask it on the first call, before you say anything about the job, and write the answer yourself in the candidate's words. A dropdown on a form gets "other" or whatever sits at the top. Try: "Before I tell you about the role, how did you hear of it?" If the answer is "a friend", ask which friend works here, and note the name. That one follow-up shows you which of your current staff carry your name outside, and they are the people to thank.
When does a reason become a decision?
Act when the same reason appears twice, not once. One leaver who found the shift long may be unlucky. Two have read a poster that does not match the job. Change the poster or the shift, write the change in the register's margin with the date, and read the next rows against it.
What should stay out of the register
Do not turn it into a scorecard for the person who keeps it. A leaver's real reason is sometimes uncomfortable, and an owner who punishes the messenger gets blank reasons. Keep names and reasons on one private sheet. If you share results with managers, share counts only.
Back-fill your last ten hires
Fill rows for your last ten joiners from memory and the attendance records, with the source and the exit reason. The gaps you find are the first result. The employer branding hub covers the rest of the work. If you want a second pair of eyes on what the sheet says, campaign analytics and performance reporting is the service nearest to it, and you can tell us what you need.
Related services
