Two days before the deadline and the sample is thin
Your state-of-trade survey matters. Government officials quote it, journalists ring about it, and your policy team uses it in every consultation response. Each quarter the research officer sends the questionnaire to the member mailing list, waits, sends a reminder, waits, and then starts phoning the firms who always used to respond.
Some of the regulars have stopped. A few send a spreadsheet with half the columns blank. One large member sends figures in a different unit from everyone else. By the close date you have enough to publish, just, but the regional split is too thin to say anything about the North East, and the small-firm category is mostly three firms.
Next quarter it starts again, with the same mailing list and the same generic reminder.
Why response rates drift down
Members want the survey to exist. What stops them returning it is friction and a lack of anything in it for them.
- The survey link is the same for everyone, so each firm starts from a blank form and has to look up last quarter's figures.
- It goes to whoever is on the general mailing list, rarely the person who holds the numbers.
- Reminders go to everybody, including firms that already responded.
- Nobody sees their own figures against the sector afterwards, only the press release.
- The research team cannot see who has started but not finished, so chasing is blind.
What a weak sample costs
Your credibility, mainly. A survey quoted in a consultation response or a select committee submission needs a sample that stands up to a sceptical question. Thin cells mean you cannot report some breakdowns at all, and the ones you can report move around more than they should from quarter to quarter.
Then there is staff time: the phone round, the reformatting, the manual reminders. Research officers spend the collection window chasing rather than analysing.
A survey flow built for member firms
We build the collection around the member record, so each firm gets its own survey and your team sees the response picture as it forms.
- Each firm has a named survey respondent on its member record, separate from the general contact, with a deputy for holiday cover.
- Every quarter the respondent gets a personal link, with the form pre-filled from last quarter where your methodology allows, so they update rather than start over.
- Units, periods and definitions are fixed in the form, with range checks that query an answer wildly out of line with that firm's history before it is submitted.
- Reminders go only to firms that have not finished, on a schedule you set, and the research team sees a live board of complete, partial and not started by region and size band.
- A partial response is saved, so a respondent can come back to it.
- Once the quarter is published, each respondent gets a short personal report comparing their figures with the sector, using your own disclosure rules.
| Response state | What happens automatically | What the research officer does |
|---|---|---|
| Not opened | Reminder to respondent, then to deputy | Phones only the firms your sample most needs |
| Started, not finished | Gentle nudge with a resume link | Nothing unless the deadline is close |
| Submitted with a flagged value | Held for query | Checks with the firm |
| Submitted and clean | Added to the dataset | Nothing |
We can build this with an off-the-shelf survey tool underneath, such as Microsoft Forms or a hosted survey product, or as a custom form. The choice depends on your data handling rules and how much pre-filling you need.
The collection window, reworked
The research officer opens the board on day one and sees responses arriving. Calls go to the firms whose absence would hurt the sample, not to everyone. Returns arrive in the right units because the form will not accept anything else. When the analysis starts, it starts on a dataset, not a pile of spreadsheets.
And the members who took part get something back with their name on it, which gives them a reason to answer next quarter.
Check your own survey round
- The survey link is identical for every member.
- Reminders go to firms that have already responded.
- Your research officer spends the deadline week on the phone.
- Some breakdowns cannot be published because the cells are too small.
- Respondents never see their own figures against the sector.