The 2026 marketing benchmarks report:every number sourced, India and US side by side.

Two Indian counts, and why they differ. The FICCI-EY Media and Entertainment report, released on 24 March 2026, put total Indian advertising at ₹1.5 trillion (₹1,50,000 crore) in 2025, up about 13%, with digital at ₹94,700 crore. The dentsu Digital Advertising Report 2026, released on 2 February 2026, put total ad spend at ₹1,21,339 crore and digital at ₹71,621 crore. The gap is mostly scope: FICCI-EY counts more of the long tail, including ₹36,300 crore of digital advertising from SMEs and small advertisers, while dentsu's estimate is built around media spend by advertisers it tracks. Neither is wrong. Use one source consistently, and do not compare a FICCI-EY total with a dentsu share.

What the counts agree on. Digital is now well over half of Indian advertising on both counts. Commerce media, meaning ads sold by Amazon, Flipkart, quick-commerce apps and other retailers on their own sites, is the fastest-growing format: FICCI-EY puts e-commerce and point-of-sale advertising at ₹22,000 crore, up 50%, and dentsu puts e-retail ad spend at ₹17,601 crore, up 55.86%, or about a quarter of all digital spend. Search and social together take 64% of digital advertising in the FICCI-EY figures.

Putting the two markets side by side. At ₹95.7485 to the dollar, FICCI-EY's ₹94,700 crore of Indian digital advertising is about $9.9 billion, roughly one-thirtieth of the US figure. Paid search shows the gap more sharply: dentsu's ₹16,581 crore for Indian search is about $1.7 billion, against $114.2 billion in the US. That ratio says more about price per click and order values than about how many people search. India has far more internet users than the US (IAMAI and Kantar counted 886 million active users in 2024), but each click and each order is worth less, so auctions clear at lower prices.

What it means for planning. Two shifts show up in both markets. First, commerce media is taking budget that used to sit in search and display, because it lets brands reach shoppers at the point of purchase with the retailer's own data. If you sell on marketplaces or quick commerce, a line for retail media belongs in the plan, measured on incremental sales rather than platform-reported ROAS. Second, video and social keep gaining at the expense of linear formats. Neither trend means search is shrinking; it means search's share is being diluted by formats growing faster.

What the vertical mix tells a smaller advertiser. FMCG and e-commerce together account for nearly half of Indian ad spend on dentsu's count, and more than half of digital. Those two verticals bid for the same consumer audiences on Meta, YouTube and Google that a D2C brand, an edtech company or a consumer fintech needs to reach. Their seasonal peaks, especially the festive quarter, raise auction prices for everyone else. A smaller consumer advertiser planning a festive push should expect to pay more per impression and per click than in the quieter months, and either plan for it or deliberately concentrate spend outside the peak weeks.

B2B advertisers are partly insulated: job-title targeting on LinkedIn and specialised search terms on Google are not contested by FMCG. But B2B advertisers using Meta for SMB audiences share inventory with consumer brands and feel the same seasonality. The seasonal ad CPM predictor models the effect on a plan.

Dreamdata, which sells B2B attribution software, measured return on ad spend against closed-won revenue over 12 months for its customers and found LinkedIn ahead of Google Search and Meta. It also counted an average of 10 stakeholders in a B2B buying journey, up from 6.8. Treat the ranking with care: it is one vendor's customers, mostly B2B software, with that vendor's attribution model. But it supports a point that cost-per-lead comparisons miss. A $202 lead that reaches the right job title at a target account can be worth more than ten $27 leads that never reach sales.

Signal loss differs between India and the US. StatCounter put iOS at 53.59% of mobile operating systems in the US in September 2026 and only 6.05% in India. Apple's privacy controls therefore remove far more conversion signal from US campaigns than from Indian ones. Indian advertisers still lose signal to browsers, ad blockers and broken tracking, but the main cause of poor Meta optimisation in India is more often missing or duplicated events than iOS. Meta's own help centre says an ad set needs about 50 optimisation events in a week to leave the learning phase; for many Indian B2B advertisers, that is the binding constraint, not the cost per click.

Reach in India. DataReportal's Digital 2026 report counted 481 million users reachable with Instagram ads and 403 million with Facebook ads in India, the largest ad audiences on either platform anywhere. LinkedIn says it has more than 160 million members in India, its second-largest market. Scale is not the constraint for Indian advertisers on any of the three platforms. The constraints are signal quality, creative volume and the speed of sales follow-up.

Attribution windows. Meta reports conversions on a default of 7-day click and 1-day view. A view-through conversion counts when someone saw an ad and later converted without clicking, which flatters upper-funnel campaigns. When comparing Meta with Google or LinkedIn, put all three on the same window, or better, compare each against backend revenue in your CRM.

Measurement is the benchmark that matters most on Meta. Every figure in the tables above assumes the platform knows which clicks turned into leads or sales. If your pixel fires twice, your server events lack an event_id, or your qualified-lead events never reach Meta from the CRM, the platform optimises towards the wrong people and every cost metric drifts. For Indian advertisers, where Android dominates and iOS signal loss is limited, these setup errors are usually a larger cause of weak results than anything in the auction.

For B2B advertisers, the most valuable step is sending a second, later event back to Meta: a lead that sales has qualified, or a deal that closed. Meta's Conversion Leads integration exists for this purpose. It lets the platform learn the difference between a form fill and a buyer, which a cost-per-lead benchmark cannot capture. The same principle applies on Google (offline conversion imports) and LinkedIn (its Conversions API).

Planning LinkedIn audiences. LinkedIn will run an ad to as few as 300 members, but it suggests at least 50,000 to drive results and 300,000 for Sponsored Content and messaging. Account-based campaigns aimed at a few hundred target companies sit far below those numbers, so they pay higher CPMs and reach the same people often. That is a reasonable trade for a high-value account list, as long as it is planned: cap frequency deliberately, rotate creative, and judge the campaign on engagement and pipeline at the target accounts rather than on cost per click. Company list uploads accept up to 300,000 companies and need at least 300 rows, with 1,000 or more recommended.

The minimum daily budget of $10 (about ₹957) makes testing cheap, but a test at that level rarely produces enough clicks to judge conversion. At a $9.39 median cost per click, $10 a day buys about one click; a month buys about 32. A meaningful test usually needs several times the minimum.

Tools and resources for this chapter

First Page Sage, a marketing agency, compiled these rates from its client data between 2019 and 2025 and published them as its 2026 report (last updated 23 December 2025). Two cautions apply. MQL definitions vary widely: a company that calls every form fill an MQL will convert far fewer to SQL than one that requires a fit score and a buying signal. And this measures MQLs that became SQLs, which is stricter than "sales accepted". A rate far above these ranges usually means a strict MQL definition, not unusual sales skill; a rate far below usually means marketing is counting leads sales never wanted.

Response time. The Harvard Business Review audit of 2,241 US companies, published in 2011, remains the most cited primary study on lead response. 37% replied to a web lead within an hour, 23% never replied, and the average response time was 42 hours. Firms that tried to contact a lead within an hour were nearly seven times as likely to qualify it as those that waited even one hour longer. The study is old and US-only, but nothing published since with a comparable method contradicts it, and it matches what any sales manager sees in a CRM: interest decays in hours, not days.

Buyers decide before they talk to you. 6sense's 2025 B2B Buyer Experience Report found buyers first contacted sellers 61% of the way through their journey, earlier than the 69% the previous edition reported, and that the winning vendor was already on the day-one shortlist in 95% of deals. Add the B2B Institute's 95:5 rule, reported by Marketing Week and based on Professor John Dawes' work, that only about 5% of B2B buyers are in market at any given time, and the implication is uncomfortable for lead-generation-only plans. Most of the work that decides who wins happens before a form is filled, among people who are not buying yet.

Forecasting is where funnel problems surface. Xactly's 2024 Sales Forecasting Benchmark Report, as reported by Agile Brand Guide, found that four in five sales and finance leaders had missed a quarterly forecast in the past year, 52% of sales leaders said their forecasts were off by 10% or more, and 60% were unsure where their pipeline data came from. Gartner found in 2020 that only 45% of sales leaders and sellers had high confidence in their forecast accuracy. A forecast is only as good as stage definitions and the discipline with which reps update them.

Measuring the funnel honestly. Three practices make the benchmarks in this chapter usable. First, measure conversion between stages on cohorts (all MQLs created in March, and what became of them) rather than on activity in a period (MQLs in March against SQLs in March), because the second mixes leads of different ages. Second, record time stamps for every stage change so cycle length and response time come from the CRM, not from memory. Third, require a closed-lost reason from a fixed list; without it, a falling win rate cannot be diagnosed.

India. No Indian dataset on MQL-to-SQL rates, response times or win rates with a published method was found. Indian B2B companies frequently receive leads through IndiaMART, JustDial, WhatsApp and phone calls as well as web forms, and those sources rarely reach the CRM with time stamps. Before comparing response time with the HBR figures, make sure every source is captured, timed and routed automatically.

Tools and resources for this chapter

Growth rates fall as companies get bigger, which is why a single "good growth rate" is meaningless without revenue size. 75% growth below $1 million ARR is a small absolute number; 15% above $50 million is a large one. SaaS Capital's figures also show the gap between bootstrapped and equity-backed companies is smaller than many founders assume, five percentage points at the median, while equity-backed companies spend twice as much on marketing (chapter 3).

Efficiency metrics. Four ratios dominate board conversations. The Rule of 40 adds revenue growth to profit margin and asks whether the total reaches 40%. The magic number divides new annualised revenue in a period by sales and marketing spend in the previous one. The burn multiple divides net cash burn by net new ARR. CAC payback counts the months of gross margin it takes to earn back the cost of acquiring a customer. The published ranges for each are below.

Product-market fit before efficiency. None of these ratios mean much before a product has found its market. The best-known quantitative test is Sean Ellis's: ask users how they would feel if they could no longer use the product, and look for at least 40% answering "very disappointed". Superhuman's founder described in First Round Review starting at 22%, reaching 33% after narrowing to its best-fit users, and 58% about three quarters later. The lesson is in the segmenting: efficiency improves fastest when a company stops selling to customers it cannot keep.

India. Bain & Company put Indian SaaS ARR at $12 to $13 billion in 2022. Indian SaaS companies selling globally are benchmarked by investors against the same US-centric figures above; there is no separate Indian dataset for Rule of 40 or magic number with a published method. Two local differences matter in practice: lower salary costs can make early efficiency look better than it is if the company has not yet paid for senior go-to-market hires in its target market, and rupee-denominated revenue from Indian SMBs usually carries lower contract values and higher churn than the samples here.

Reading the ratios together. Each efficiency metric answers a different question, which is why a single one can mislead. The magic number asks whether sales and marketing spend is buying revenue. CAC payback asks how long that revenue takes to repay the spend, after cost of delivery. The burn multiple asks whether the whole company's cash consumption is buying growth. The Rule of 40 asks whether growth and profitability together are strong enough. A company with a strong magic number but long payback is probably buying low-margin revenue; one with a good burn multiple but weak Rule of 40 is probably spending heavily on something other than growth. The diagnosis lives in the disagreement between the ratios.

Definitions differ. Benchmarkit does not state the margin basis for its Rule of 40 figures; Aventis uses profit margin for public companies; many private companies use EBITDA margin or free cash flow margin. Magic number is sometimes calculated on gross-margin-adjusted revenue. Before comparing with any figure here, state which definition you use and use it every quarter.

Tools and resources for this chapter

Unicommerce's April 2026 India D2C report draws on transaction data from its Uniware and Shipway platforms, covering more than 6,000 D2C brands and more than 410 million shipments. Unicommerce sells software that addresses RTO, and these figures are platform-wide rather than category-specific, but it is the largest published dataset on the problem. The festive peak matters because it coincides with the highest ad spend of the year: brands buy the most expensive clicks of the year to generate orders of which, at the peak, nearly two in five came back.

Payments are shifting. Unicommerce's monthly benchmark showed prepaid orders rising from 51% to 55% of D2C orders between May and June 2026. UPI processed 24.07 billion transactions worth ₹29.37 lakh crore in September 2026, up 23% on the year, according to NPCI figures reported by DD News. A higher prepaid share lowers RTO. Common ways to raise it without losing the customers who need COD include small prepaid incentives at checkout, COD confirmation on WhatsApp before dispatch, and partial prepayment on high-value orders.

McKinsey projects India's D2C channel growing about 38% a year to 2030, as reported by Apparel Resources, and Bain & Company counted more than 270 million Indians shopping online in 2024. The headline growth hides two pressures. Quick commerce is taking a growing share of everyday categories at platform-set margins with little customer data for the brand. And the half of orders coming from Tier II and III cities carries higher COD shares and higher RTO than metro orders, which raises the real cost of each delivered order.

Break-even ROAS in India is higher than the textbook says. The textbook break-even is 1 ÷ gross margin: a 60% margin needs 1.67× ROAS. An Indian D2C brand also pays forward shipping on every order, return shipping on RTO, gateway fees on prepaid orders and COD handling fees, and earns nothing on orders that come back. The chart below, from this site's worked example, shows how quickly break-even rises with RTO at a 50% gross margin.

The example shows why the same ROAS can mean profit for one Indian brand and loss for another. A brand selling mostly prepaid to metro customers might break even near 1.9×; a brand selling the same product mostly on COD to Tier II and III cities might need 2.4× or more. When an agency or a peer quotes a "good ROAS", the first question is what their COD mix and RTO rate are.

Repeat purchase changes the answer. First-order break-even is the right test for a single sale. Brands with strong repeat rates can afford to acquire below first-order break-even and earn the margin back on later orders, as long as they measure repeat behaviour by cohort and do not assume it. The MER ranges below show US brands running below break-even deliberately for the same reason. The risk in India is that RTO-heavy first orders bring in customers who never complete a delivery at all, so cohort repeat rates should be measured on delivered customers, not placed ones.

Where to sell: website, marketplace or quick commerce. The Diwali 2025 figures show orders growing fastest on quick commerce (+120%) and on brand websites (+33%), and slowest on marketplaces (+8%). Each channel has different economics. A brand website carries the full cost of acquisition, shipping and RTO but keeps the customer data and the repeat relationship. Marketplaces bring demand and handle much of the logistics in exchange for fees and limited customer data. Quick commerce brings very fast delivery to dense urban demand at margins and terms the platform sets, with almost no customer data for the brand.

The benchmark that matters for the channel decision is contribution per order after all channel costs, compared across channels, together with what each channel does for repeat purchase. A brand that measures only revenue by channel will tend to over-invest in whichever channel reports the most sales, which is rarely the one that earns the most.

Tools and resources for this chapter

Why CRM data decays. Every job change leaves a contact record with a dead email, an old title and the wrong company. The US Bureau of Labor Statistics put median tenure with a current employer at 4.1 years in January 2026, and its JOLTS survey recorded 3.2% of US workers separating from their jobs in August 2026 alone. A B2B database that is not refreshed loses accuracy every month, quietly, and the cost shows up as bounced sequences, misrouted leads and forecasts built on stale accounts. No equivalent Indian tenure statistic with a published method was found; treat the US figures as an indication of how quickly any B2B contact list goes stale.

Gartner's $12.9 million figure is an average across large organisations and is not specific to CRM, so treat it as an indication of scale rather than a cost to quote for a 50-person company. The more useful Gartner finding is the 59%: most organisations do not know how good or bad their data is, which means nobody owns fixing it.

Choosing a measurement approach by size. Most companies in this report's audience do not have two years of clean weekly data or the budget to run regular lift tests, so platform attribution is the practical starting point. The discipline is to treat platform numbers as each platform's opinion, reconcile them monthly against revenue in the CRM or store, and run a simple holdout test (pausing a channel in some regions or for some weeks) when a large budget decision depends on the answer. As spend grows and history accumulates, MMM becomes feasible; open-source tools such as Meta's Robyn make it cheaper than it used to be, but they still need good data in.

RevOps. Gartner predicted in 2021 that 75% of the highest-growth companies would deploy a revenue operations model by 2025, and LinkedIn ranked Head of Revenue Operations first on its 2023 US Jobs on the Rise list, as reported by Fortune. The prediction's deadline has passed without a published follow-up measuring it, so it is best read as evidence of direction rather than adoption. What RevOps does in practice is take ownership of the data, stage definitions, routing and reporting that the previous chapters show leaking: lead response time, MQL definitions, forecast accuracy and retention reporting.

What CRM tools cost in India. List prices change often and discounts are common, but they set the starting point for any business case. HubSpot reported 306,446 customers worldwide at 30 June 2026, up 14% on the year. The table below gives Indian list prices checked on 17 October 2026.

Choosing a CRM on cost in India. The list prices above show how wide the range is. For a ten-person sales team, year-one licence cost runs from about ₹70,000 on HubSpot's promotional Starter price, through ₹1.68 lakh on Zoho CRM Professional, to more than ₹10 lakh on HubSpot Sales Hub Professional once its required onboarding is included. The cheaper tools are not worse in general; they are built for different jobs. The deciding questions are which features the sales process actually needs (sequences, forecasting, territory rules, WhatsApp, Indian lead sources such as IndiaMART), where the data must be stored, and who will configure and maintain the system.

Tools and resources for this chapter

Two things make salary comparisons between the markets misleading if taken at face value. First, a US salary is only part of the cost of employment: the BLS put benefits at 30.0% of total compensation in US private industry in June 2026. Second, Indian salary-site averages mix very different employers, from small agencies to large consumer companies, and the senior-role samples are small. For budgeting a specific hire, current job listings and recruiter quotes in your city and sector are a better guide than any average.

Turnover is a cost too. Spencer Stuart's 2026 snapshot put average CMO tenure at S&P 500 companies at 4.1 years against 5.0 for the wider C-suite. Every change of marketing leader brings recruiting cost, a ramp period and, often, a change of strategy and agencies. Smaller companies feel this more, because one senior departure can stall the whole function.

What the comparison leaves out. A full-time hire carries costs beyond CTC: recruiting (often a percentage of first-year pay through a search firm), notice periods that delay joining, ESOPs, and the cost of a wrong hire, which at senior level is measured in quarters of lost momentum. A fractional arrangement carries different risks: fewer hours, competing priorities and a dependence on the team or agency underneath to execute. The right answer usually depends on how much execution capacity already exists. A company with a capable team and agencies but no senior direction gains most from fractional leadership; a company with no team needs execution capacity first, and senior direction later.

Tools and resources for this chapter

Why video and commerce media are winning budget. Both formats solve a problem advertisers have with older channels. Video, especially short-form on social platforms and connected TV, reaches audiences that have left linear television: EY's figures show linear TV ad revenue in India falling 10% in 2025 while connected TV advertising rose 42%, with about 40 million connected TV sets in Indian homes, up from 30 million in 2024. Commerce media reaches shoppers at the moment of purchase, using the retailer's own data on what people search for and buy, which is exactly the signal that browser privacy changes have taken away from open-web advertising.

What it means for Indian brands. For any brand selling on Amazon, Flipkart or quick-commerce apps, retail media is no longer optional: it decides visibility in search results inside the store. The measurement trap is the same as with platform ROAS elsewhere. A retailer reports sales that followed an ad, including sales that would have happened anyway from people already searching for your brand. Judge retail media on incremental sales, using periods or products where ads were paused, and on new-to-brand customers rather than total attributed revenue.

Programmatic keeps growing. Dentsu put programmatic at 42% of Indian digital spend in 2025, and FICCI-EY expects programmatic buying to exceed 75% of non-premium inventory by 2028. In the US, IAB counted $162.4 billion of programmatic revenue. Automated buying is now the default for display, video and connected TV, which makes first-party data and conversion signals the main thing an advertiser can control.

Tools and resources for this chapter

Three definitional traps. First, "conversion rate" in an ad benchmark usually means clicks to leads, while on a website it usually means sessions to conversions; the two are not comparable. Second, "churn" can mean customers lost or revenue lost, monthly or annually; a 5% monthly customer churn is not the same as 5% annual revenue churn, and the difference is enormous. Third, "CAC" can be paid-media cost only or fully loaded with salaries and tools; investor benchmarks assume fully loaded. Before comparing your number with any benchmark, check all three.