Almost every review-collection system has the same bias. It asks by email or SMS, so the people who respond are the ones who were sufficiently pleased to bother. The unhappy customer does not reply — they simply do not buy again, and occasionally they post about it somewhere public.
That means your review score is measuring your happy customers' enthusiasm rather than your product's quality, and your first warning about a real problem arrives as a one-star review.
What a call does differently
It reaches the silent middle. A customer who was mildly disappointed will not compose an email, but will say "actually the packaging was damaged" when asked directly, because it takes four seconds and someone is listening.
That is the group worth hearing from. The delighted customers will tell you anyway; the furious ones already have. It is the mildly disappointed who churn quietly and in volume.
Two to three days after delivery
Long enough that they have opened and used the product. Short enough that it is fresh.
Same-day is too early — they may not have opened the parcel. A week later, the detail has gone and you get a vague "it was fine".
Ask one question
The single biggest mistake is a survey. Five questions on a rating scale over the phone is unbearable, and people will agree with anything to get off the call — which produces data that is worse than none.
Ask one open question and let them talk.
Do not offer compensation on the call
The instinct when someone reports a problem is to fix it immediately with a coupon. It is a mistake for two reasons.
First, it teaches customers that the feedback call produces a discount, and the quality of your feedback collapses within weeks. Second, the appropriate remedy depends on what happened, what it cost and who the customer is — that is a human judgement.
Capture the problem and get a person to call the same day. The customer feels heard, which is most of what they wanted, and your resolution stays proportionate.
"An honest one"
Asking specifically for a positive review is against the policies of every major review platform and, on some, grounds for removing your reviews entirely.
Asking for an honest review from a customer who has just told you they are happy produces a positive review anyway, without the risk. The selection has already happened in the conversation.
What it is worth
Under a minute at ₹4–5. For a store shipping 3,000 orders a month, calling every one is roughly ₹12,000 — which is not nothing, so most stores should sample rather than call everyone.
Sensible sampling: all first-time customers, all orders above a value threshold, and everything in a category you have just changed. First-time customers are the highest-value segment here, because whether they order again is decided in the first two weeks and a good recovery call at that moment is disproportionately effective.
The by-product is the real return
Counted, categorised feedback in customers' own words. Most stores discover their complaints cluster on two or three specific things — a courier that damages parcels in one region, a size chart that runs small, a product photographed in a misleading light.
None of those show up in a star rating. All of them are fixable, and each fix improves every future order rather than just recovering one customer. Running phone surveys properly goes further into the method, and the e-commerce page covers the other calls worth making.