Effort predicts loyalty better than delight
Most chat teams measure satisfaction, and satisfaction is worth measuring. But it answers a
question about feeling rather than difficulty, and the two come apart more often than you would
expect. A customer can rate a chat five stars because the agent was lovely, after spending eleven
minutes and three repeated explanations getting to an answer that should have taken one.
Effort is the measure of how hard the customer had to work. It is the count of times they
repeated themselves, the number of places they were sent, the questions they had to answer that you
could have known already. Teams that track it tend to find that reducing effort moves repeat contacts
and retention more reliably than pushing an already-good satisfaction score higher.
The practical argument for measuring it is simpler still: effort is actionable. A three-star
rating tells you someone was unhappy. A high-effort chat tells you exactly which step made it hard,
and that step is usually something you can change.
The question, and when to ask it
The standard phrasing asks how much effort the customer had to put in to get their issue handled,
answered on a scale from very low to very high. Keep whatever wording you choose identical over time
— a rephrased question produces a new baseline, and you lose the ability to compare.
Ask immediately after the conversation ends, in the chat window, while the experience is intact.
An emailed survey the next morning measures memory, and memory of a support interaction compresses
toward whatever the emotional peak was.
One tap beats a form here even more than it does for satisfaction, because asking someone to fill
in a survey about how much work you made them do is its own small irony. If you add a free-text box,
make it optional and ask what would have made it easier, which produces far more usable answers than
a blank comment field.
Where effort hides in a chat
Four patterns account for most of it, and all four are visible without any survey at all.
Repetition. The customer explains the problem to the widget, then again to the
first agent, then again to whoever the chat was transferred to. Every retelling is effort you
imposed, and it is the single most common complaint in chat transcripts.
Redirection. Being told to email a different address, call a number, or fill in a
form is effort even when the destination is correct. Each handoff is a point where a portion of
people give up entirely.
Waiting without information. A silence with no explanation costs more than a
longer wait that was narrated. The customer is doing the work of wondering whether anyone is still
there.
Being asked for things you already hold. Order numbers, plan details and account
identifiers that are already visible to the agent, or already in the pre-chat form, are the most
irritating category because the customer can tell it was avoidable.
Reading effort without a survey
Response rates on any in-chat question are partial, so the survey is a sample rather than a
census. The full picture lives in the conversations themselves, and a structured
transcript review will find effort faster than a dashboard.
Scan for a short list of phrases. As I said, like I mentioned and again
mark repetition. I already marks a customer being asked for something they provided.
Is anyone there marks an unnarrated silence. Count the number of turns before the customer
got their first substantive answer, and count transfers per conversation. None of that needs
tooling more sophisticated than a search box and a tally.
Then look at the chats that came back. A returning conversation about the same issue within a few
days is a high-effort chat by definition, whatever it was rated at the time.
What to do with a high-effort chat
Treat effort as a process signal rather than an agent signal. Almost everything on the list above
is caused by something structural: routing that sends people to the wrong place, a pre-chat form that
collects information nobody reads, a policy that forces a transfer, a knowledge gap that makes agents
hedge and check.
Coaching an agent for a high-effort chat they inherited is both unfair and ineffective. Fixing the
routing rule that caused the transfer removes the effort from every future conversation.
Pick one cause per month rather than attacking the whole list. The repetition problem alone
— making sure context travels with a transfer — is usually the largest single reduction
available, and it is a configuration change rather than a training programme.
What to measure
Track the effort score itself as a trend, never as a single number, and always alongside its
response rate so you know how thin the sample is. Segment it before you average it: effort is usually
concentrated in a few issue types rather than spread evenly.
Pair it with two counts that need no survey at all — transfers per conversation and repeat
contacts within seven days. Those two move for real reasons and are hard to game.
Watch the relationship between effort and satisfaction rather than either alone. Chats rated
satisfied but high-effort are your best improvement list, because the customer liked the person and
still had to work too hard. That gap is where the process problems live, and closing it is the
cheapest quality gain most teams have available.