“It’s much easier to add a question to the survey,” T-Mobile’s Carly Rintisch told CX Dive in April. She was describing why teams relied on questionnaires despite having access to records of customer transactions and interactions.[1]
Her team changed that approach. One survey shrank from more than 30 questions to nine, cutting the time to complete it from roughly ten minutes to two. The team stopped asking customers for facts the company already held.[2]
For another business making that change, I would attach a condition: every question it still needs answered must have an owner inside the company. Removing the survey item is only half the job.
A survey becomes a substitute for internal investigation when customers are asked to piece together facts the business has recorded but has not connected. The customer becomes its data integration layer, reconstructing the history because the company has left the records apart.
Adding a question avoids the internal work of finding those records and making them usable. It leaves the customer to do the explaining.
At T-Mobile, changing the approach involved more than editing the questionnaire. Rintisch described working with other departments and vice presidents to get them using the behavioral data already available.[1] That is where I would expect a survey review to lead: into the teams requesting the information, with a decision about how they will obtain it.
Customer feedback still had a specific job. Rintisch described customers struggling to find information on T-Mobile’s website. Follow-up questions revealed they were often looking for benefits activation, giving the insights team a concrete issue to take back to the people responsible for the site.[1]
That question earned its place. A record of pages visited does not, by itself, establish what someone was trying to find. Asking the customer about their purpose gives the business something it cannot safely infer from the clicks.
The same caution applies to records that appear complete. A case marked “resolved” is not proof the customer’s problem was solved. Customers need room to correct the company’s account, including problems its systems never captured. The point is to avoid asking them to compensate for internal gaps, not to treat company data as unquestionable.
Before sending the next survey, review each question with the team that needs its answer. Ask what decision the answer supports and whether usable information already exists. A field somewhere in a database is not enough. Someone must be able to retrieve the right record, connect it to the experience being reviewed and check whether it answers the question.
Where the information is missing, asking the customer remains reasonable. Where it exists but nobody takes responsibility for using it, the next action belongs inside the business.
T-Mobile’s shorter survey followed a change in how the team used existing records.[2] Copying the question count would miss the lesson. For every item removed, management should know whether the answer is no longer needed or who will now supply it.
Your customer should not have to reconstruct your records because nobody inside the company owns that work.