Nutrition

Protein quality matters alongside protein quantity. Cereals and pulses supply different balances of amino acids, so each can supply what the other lacks. The traditional combination of dal with rice or roti has a nutritional purpose as well as a place in everyday meals (Swaminathan, Vaz & Kurpad, British Journal of Nutrition, 2012).

That everyday food base is visible in the survey: dal, chana or rajma is the most widely named protein source, selected in four in five answers, while protein powders sit below all five named foods. Other choices vary sharply by age, income and city. The page starts with how often people eat out or order in, then examines the protein sources and diets they name. These answers describe food habits, not nutrient intake: the survey did not record portions or ingredients, so it cannot establish whether a diet is adequate.

23.5% eat out or order in three or more times a week, and 21.7% never do#

All five answers are shown and together they account for every one of the 7,614 responses to this question. Frequent on this page means the three bands at the bottom of that table, so three or more occasions a week: 1,793 responses, or 23.5%. Another 8.5% of answers reported five or more.

Most answers sit at the low end. Three in four, 76.5% of them, report never or once to twice a week, and the frequent group is about one in four.

For scale, NCAER put the gross value of output of India's food-delivery platform sector at about ₹1.2 lakh crore in 2023–24, roughly 0.21% of national output, in a study sponsored by Prosus (Bhandari and Sahu, NCAER, 2025). Two-thirds of that is the value of food orders placed with restaurants, not platform revenue.

Eating out three or more times a week falls from 27.9% in the 20s to 10.9% in the 50s#

These four bands cover 7,552 of the 7,614 answers. The other 62 fall outside 20 to 59 or have no usable age.

The share in the twenties is about two and a half times the share in the fifties, 27.9% against 10.9%, and it falls at every step in between. The oldest band rests on 229 answers and 25 selections.

The familiar explanations — households form, kitchens get bigger, cooking becomes routine — are plausible, and this survey tested none of them.

The previous edition found the same shape on a smaller sample of 3,437 responses, grouped by generation rather than by decade: ages 22 to 30 at 27.2%, 31 to 40 at 19.2% and 41 to 55 at 18.2%. Those are three groups against this year's four, and the oldest of them spans fifteen years where 2026 uses two separate decade bands, so the individual numbers are not a like-for-like comparison. The direction is the same in both editions.

Among the seven metros, about 30% in Chennai and Bengaluru eat out three or more times a week, against 18.9% in Pune#

The table shows seven metros with about 250 or more answers, not all city groups. WHI 2025 gives no city sample sizes; differences between editions are sample differences, not city trends.

Frequent eating out is about 60% more common in the Chennai and Bengaluru answers than in the Pune answers. That 11.5-point spread across the seven metros is far wider than the 1.8 points between the sexes, at 24.1% of 5,234 male answers against 22.3% of 2,351 female answers.

The sexes were further apart in the previous edition. WHI 2025 reported frequent eating out at 31.2% of male responses and 24.1% of female responses on its 3,437-response sample, a gap of 7.1 points against 1.8 points here, with the male share carrying almost all of the difference. That page does not publish its per-sex sample sizes, so this is a comparison of two published shares rather than a tracked change.

Of the five metros both editions publish, Hyderabad was the lowest in 2025 at 22.1% against Pune's 22.2%, and Pune is the lowest in 2026. Bengaluru is the only one of the five reading higher this year. Neither is a statement about behaviour changing.

Work pattern separates the answers too, though less sharply than age. Frequent eating out runs 26.9% among 1,618 hybrid answers, 23.7% among 4,653 full-time office answers and 19.1% among 1,341 work-from-home answers, a spread of 7.8 points. The Work Model page carries the wider set of cuts by work pattern.

Pulses were named in four in five protein answers, whey or casein in about one in eight#

The protein question is multi-select. A respondent could tick every source they use, so the shares overlap, the column does not sum to 100%, and a source that is not ticked is not proof that someone never eats it.

Four in five answers name pulses, and pulses lead the table in both editions. Below them the table falls away. Fish, the least-selected of the five foods, is named in 36.3% of answers against 12.1% for whey or casein, the most-selected of the two powders — about three times as often. The two powders sit below all five named food sources.

Set against WHI 2025: dal was 82% of 3,437 responses then and 80.7% of 7,632 now, and whey or casein 11.8% then and 12.1% now, the smallest movement of the six sources with a 2025 counterpart. Eggs moved most, from 55.2% to 63.5%, with chicken from 52.9% to 56.5% and fish from 33.3% to 36.3%. Two of the six fell — dal, and plant-protein powder, which fell furthest, from 10.3% to 7.7%. The 2025 shares are published figures on a sample less than half the size, so these are two survey samples rather than a tracked population.

Whey selection climbs with income and falls with age#

The seven named bands, the 580 who preferred not to say and three answers with no income entry account for all 7,632 responses to the protein question.

Selection almost doubles from the lowest band to the highest and rises at every one of the six steps between them, although the step from ₹6–10 lakh to ₹10–15 lakh is only 0.1 point. In absolute terms the gradient is shallow, 9.2 percentage points end to end, and it holds across all seven ordered bands.

The survey does not say why. Price, habit and the circles a person moves in are all candidates and none of them was asked about, and nothing here shows whether anyone in any band is short of protein.

Selection runs the other way with age.

These four bands cover 7,571 of the 7,632 protein answers. The other 61 fall outside 20 to 59 or have no usable age. Selection in the twenties is more than three times the rate in the fifties, a fall of 10.7 points across the four bands, and the oldest band rests on 229 answers and 10 selections. Men named whey or casein more often than women, 13.3% of 5,252 male answers against 9.3% of 2,351 female answers, a gap of 4.0 points.

Whey or casein is one of eight protein-source options. Reported supplement use is covered on the Supplementation page.

Whey or casein was selected by 11.8% of responses in 2025 and 12.1% now. Broader income cuts are on the Income & Job Role page.

Three in five responses declared non-vegetarian, a third vegetarian#

All three answers are shown and together they account for every one of the 7,669 diet declarations. Eggetarian is a category the survey offered and a term used mainly in India. It means someone who eats eggs but not meat or fish.

The 2026 sample declared vegetarian less often than the 2025 sample did, 32.7% against 36.3%, with non-vegetarian at 58.2% against 56.9%. These are two differently sized samples answering the same question, so the difference is between the samples and not a change in what anyone eats.

A declared label is not a measure of intake. The survey asked which diet someone follows, not how much of any food they eat, so no answer here gives an amount of animal-source food or an amount of any nutrient.

Declared vegetarian runs from 48.7% in Delhi NCR to 14.2% in Kolkata#

Almost half the Delhi NCR declarations are vegetarian. Hyderabad and Chennai are level at 16.1%, a third of the Delhi NCR rate and 32.6 points below it, and Kolkata is lower again at 14.2%. Across these seven metros the spread is 34.5 points, three times the 11.5 points that separate the same seven cities on frequent eating out.

The comparison covers the same seven metros, not the highest and lowest of all city groups. Survey diet answers are not linked to individual laboratory results.

Fish selection runs from 69.0% in Kolkata to 19.7% in Delhi NCR#

Kolkata names fish 3.5 times as often as Delhi NCR, Chennai 2.6 times and Bengaluru 1.9 times. The 49.3 points between the two ends is the widest of the three city contrasts this page shows across these seven metros, ahead of 34.5 points on declared vegetarian diet and 11.5 on frequent eating out.

The two rankings point opposite ways at their ends. Of these seven metros, Delhi NCR has the highest declared vegetarian share and the lowest fish selection, and Kolkata has the lowest vegetarian share and the highest fish selection. In between, the two orderings are close but not identical, because Mumbai and Bengaluru change places. There is no single national answer to where protein comes from. There are several, and they sort by city.

Delhi NCR is lowest among these seven metros, not all city groups. The question records protein sources named, not amounts eaten.