Most post-service surveys ask too much, too soon

Pew Research Center caps its online panel surveys at fifteen minutes, and reckons an ordinary question takes a respondent about ten seconds. By that arithmetic a twenty-question post-service survey is three or four minutes of work. That sounds like nothing. You are asking for it from someone in the worst week of their life, eight days after you buried their mother.

Most post-service surveys ask too much, too soon. But the usual fix — “make it shorter” — is the wrong instruction, and the research people cite for it doesn’t quite say what they think it says.

What the evidence actually says about length

Pew states the practitioner’s version plainly: “The longer a survey is, the fewer people are willing to complete it.” That matches every instinct you have about your own inbox.

The academic picture is more interesting. A systematic review and meta-analysis in Value in Health pooled twenty studies on questionnaire length and response rate. It found the association you’d expect — response rates were lower for longer questionnaires — but flagged its own result: the studies weren’t homogeneous, and “it is impossible to separate the impact of content from length of the questionnaires.”

Its conclusion is the useful sentence, and it’s not the one that gets quoted: “it is preferable to base decisions on use of instruments on the content rather than the length per se.”

Read that twice, because it reframes the whole problem. Long surveys may perform badly not because they’re long but because length is what happens when nobody decided what the survey was for. Cutting a twenty-question survey to ten by deleting the bottom half doesn’t fix anything — you still have ten questions nobody chose.

This is general survey methodology, not deathcare research. Nobody has run this study on bereaved families, and I’d be inventing a number if I told you the effect size for your callers. The direction is well-established; the magnitude for your situation isn’t known.

The counterargument, which is a good one

You can’t improve what you don’t measure. Whoever told you to survey after a service was probably right, and probably credible, and this post isn’t an argument for flying blind.

But look at what a typical post-service survey actually asks. How many people attended. Whether the arrangements went smoothly. Whether you followed up. Whether the family would recommend you.

You already know most of that. Attendance is sitting in your funeral guestbook. Whether you followed up is in your own calendar. The arrangement details are in the file. Asking a grieving family to re-report facts you already hold isn’t research — it’s asking them to do your data entry, at the worst possible moment, and then feeling good about being customer-focused.

Measure what only the family can tell you

That’s the whole reframe, and it’s what the meta-analysis is pointing at when it says decide on content.

Split your questions into two piles. In the first, everything you could answer from your own records — attendance, timings, who did what, whether the flowers arrived. In the second, the things that exist only inside the family’s experience: whether they felt looked after, whether anything was confusing, whether anyone treated them impatiently, what they wish they’d known a week earlier.

Delete the first pile. Not shorten — delete. It’s already in your building.

What’s left is usually two or three questions, and it isn’t a shorter survey. It’s a different instrument: a courtesy, aimed at the one thing you genuinely cannot get any other way. That the result happens to be short is a consequence, not the goal — which is precisely why “just make it shorter” produces worse surveys than this does.

The pile you deleted is worth a second look, incidentally. If you found yourself unable to answer some of it from your own records, that’s not a survey problem. That’s a records problem wearing a survey’s clothes, and it’s worth knowing which of those records you’re actually required to keep.

And send it later than feels natural

There is no good study telling you the right interval, and I’m not going to invent one.

What’s defensible is the reasoning. A survey sent four days after the service competes with the estate, the relatives who are still in town, and the fact that the person is still dead in a way that hasn’t become survivable yet. A response you get in that week is a response from someone being polite to you.

Later costs you almost nothing. The family’s memory of whether they felt looked after does not decay in three weeks — that impression set on the day and it will hold. And a note that arrives after everyone else has stopped calling is a different kind of contact from one that arrives while the casseroles are still coming.

One caution: whatever you send, and whenever, the rules on emailing people from a service are narrower than most funeral homes assume, and they turn on what the message is actually for. We covered exactly where that line sits — it’s worth reading before you build a follow-up sequence on top of a guest list.

The reframe, in one line

The survey isn’t a research instrument you’re administering. It’s a courtesy you’re asking a grieving person to perform for you — and pricing it honestly is what sets its length.

Two questions a family can only answer themselves, sent three weeks out, will teach you more than twenty questions sent on day four. Not because short wins. Because you finally asked them something only they knew.