What Is the ROI of Website Personalization for Ecommerce? Real Numbers (2026)

The real numbers behind website personalization ROI: McKinsey, BCG, Twilio Segment, Epsilon, controlled experiments and vendor case studies, plus a simple ROI formula for Shopify.

By Christopher Weiland, Founder of Ariven · Updated · 11 min read
Short answer

McKinsey says personalization most often drives a 10 to 15% revenue lift, with company-specific results from 5 to 25%. Randomized experiments, the only evidence that proves cause, find smaller and more specific effects: a 7.5% conversion lift from product recommendations across 184,375 users, a 31% lift in email leads from adding a name to the subject line. Your own ROI is the incremental revenue per visitor in a segment, times that segment's sessions, minus cost, and you can only know it by testing against a control group.

The ROI of website personalization is one of the most quoted and least checked numbers in ecommerce. Most "personalization statistics" pages repeat the same handful of figures, often stripped of who measured them, how, and when. This page collects the real numbers on ecommerce personalization revenue lift, conversion rate and average order value, labels each by type of evidence, and then shows how to work out your own Shopify store's ROI with a simple formula and a control group.

I opened every source below in November 2026 to check the exact figure and wording, except where noted.

The numbers, by type of evidence

Figure Who measured it Sample and year Type of evidence
Personalization "most often drives 10 to 15 percent revenue lift", with company-specific lift "spanning 5 to 25 percent" McKinsey Next in Personalization 2021 report Consultancy analysis
Faster-growing companies "drive 40 percent more of their revenue from personalization" than slower-growing ones McKinsey 2021 Consultancy analysis (correlation)
71% of consumers expect personalized interactions; 76% get frustrated when it doesn't happen McKinsey 2021 Survey of consumers
Brands using advanced personalization "are seeing revenue increase by 6% to 10%", "two to three times faster than those that don't" BCG May 2017 Consultancy analysis (correlation)
Revenue growth of retail personalization leaders is 10 percentage points higher than laggards BCG Personalization Index, November 2024 Consultancy index (correlation)
Returns on personalized offers are "three times higher than the returns from mass promotions" BCG BCG's work with large retailers, 2024 Consultancy observation
80% of business leaders say consumers spend more (38% on average) when their experience is personalized Twilio Segment 500 business decision-makers, March 2023 Survey of executives
56% of consumers say they will become repeat buyers after a personalized experience Twilio Segment 3,001 online shoppers in 12 countries, March 2023 Survey of consumers
80% of respondents are more likely to do business with a company that offers personalized experiences Epsilon 1,000 US consumers aged 18–64, 2017 (published January 2018) Survey of consumers
73% of customers expect companies to understand their unique needs and expectations Salesforce 13,020 consumers and 3,916 business buyers in 29 countries, Dec 2021–Feb 2022 Survey of consumers and buyers
Product recommendations raised final conversion rate 7.5% (views +15.3%, conversion given a view +21.6%) Lee and Hosanagar, Management Science 184,375 users at a top retailer, 2021 Controlled experiment
Recipient's name in the email subject line raised opens 20% (9.05% to 10.80%) and sales leads 31% (0.39% to 0.51%) Sahni, Wheeler and Chintagunta, Marketing Science Millions of recipients across three companies, 2018 Controlled experiment (email)
Revenue from users in the test group up 42%; $50.02 more per transaction AWS blog, Hoopologie (Magento) 50/50 split, 6 weeks, January 2020 Vendor case study (A/B tested)
18.3% revenue increase from personalized ranking on category pages AWS, Pomelo Fashion One retailer, 2020 Vendor case study (method not stated)
34% lift in revenue per session over 11 months Shoplift, Our Place One brand, a series of tests, April 2025–March 2026 Vendor case study (cumulative across tests)
AOV lifts from 5.84% to 63.78% across 21 published case studies Rebuy Brands selected by the vendor, checked November 2026 Vendor case study (no control group described)

Also watch for a common misreading. McKinsey's 40% does not say personalization adds 40% revenue. It says faster-growing companies get 40% more of their revenue from personalization than slower-growing ones. That is a comparison between two groups of companies, and it can't tell you which caused which.

Why only controlled experiments tell you causal lift

Every row in that table is real, but the rows answer different questions.

  • Surveys of consumers (McKinsey 71%, Epsilon 80%, Twilio Segment 56%, Salesforce 73%) measure what people say they want or would do. Stated intent is a poor predictor of what a shopper does on your product page, and none of these surveys measure revenue.
  • Surveys of executives (Twilio Segment's 38%) measure what business leaders believe. They are usually the people who bought the personalization program, and the survey can't check their numbers.
  • Consultancy analysis and indexes (McKinsey, BCG) compare companies that personalize well with companies that don't. Leaders also tend to have bigger budgets, better data and stronger brands. The comparison shows correlation, not cause.
  • Vendor case studies are chosen by the vendor. You see the wins, not the tests that lost. Many don't describe a control group, and some report before-and-after numbers that mix in seasonality, ad spend and promotions. A Shoplift article on loyalty programs puts the problem well: "a comparison of members and non-members mixes what the program did with who joined it." That applies to any "visitors who saw X spent more" claim.
  • Controlled experiments randomly split visitors into a group that gets the personalized experience and a control group that doesn't, over the same period. Because the split is random, the only systematic difference between the groups is the experience. That is the only design here that measures causal lift.

The experiments in the table are also the most modest numbers in it. A 7.5% conversion lift from recommendations. A subject-line name moving sales leads from 0.39% to 0.51%. That's what real effects tend to look like: meaningful, specific to one change, and far smaller than the headline vendor claims.

Experiments also show how often good ideas fail. In Harvard Business Review, Ron Kohavi and Stefan Thomke write that "at Google and Bing, only about 10% to 20% of experiments generate positive results," and at Microsoft as a whole "one-third prove effective, one-third have neutral results, and one-third have negative results." A personalized page is a hypothesis. Some will win, some won't, and some will lose money if you never check.

A simple personalization ROI formula for a Shopify store

You don't personalize "the website". You change what one segment sees. So calculate ROI one segment at a time:

Monthly incremental revenue = (RPV with personalization − RPV of control) × monthly sessions in the segment

Monthly ROI = (incremental revenue × gross margin) − monthly cost
  • RPV is revenue per visitor (or per session): revenue divided by sessions. It captures conversion rate and order value together, which is why it is a better primary metric than conversion rate alone.
  • Sessions in the segment means only the visitors who see the personalized version, not your whole store.
  • Gross margin turns revenue into profit. A revenue lift at a 30% margin is worth half what the same lift is worth at 60%.
  • Cost is the app fee plus your time to build and maintain the alternate pages.

Worked example (hypothetical numbers)

Every number in this example is invented for illustration, except the Shopify median conversion rate. It is not a prediction of what personalization will do for your store.

  • The median Shopify store converts 1.4% of sessions (Littledata, 436 stores, July 27 to October 24, 2026; see our Shopify conversion rate benchmarks). Assume your paid social first-time visitors convert at that rate. In practice new visitors often convert lower.
  • Hypothetical average order value: $60. Baseline RPV = 1.4% × $60 = $0.84.
  • Hypothetical segment size: 8,000 sessions a month from paid social.
  • Hypothetical result: a dedicated page for that segment lifts RPV 15%, to $0.966.
  • Hypothetical gross margin: 60%. Hypothetical tool cost: $99 a month.
Step Calculation Result
Incremental RPV $0.966 − $0.84 $0.126
Incremental revenue $0.126 × 8,000 sessions $1,008 / month
Incremental gross profit $1,008 × 60% $604.80 / month
Net monthly return $604.80 − $99 $505.80 / month

Two numbers matter more than the result:

  1. Break-even lift. To cover $99 a month at a 60% margin, you need $165 of incremental revenue, which is about $0.021 per session across 8,000 sessions. That's a 2.5% RPV lift. Any lift below that loses money; any lift above it pays for the tool. Work this out for your own store before you start.
  2. Time to know. Detecting a 15% relative lift on a 1.4% conversion rate takes roughly 50,000 sessions per group (a standard rule of thumb at 80% power and 5% significance; check your own numbers with Evan Miller's sample size calculator). With 8,000 segment sessions split 50/50, that's about a year. A 30% lift needs about 12,500 per group, roughly three months. RPV is noisier than conversion rate, so treat these as minimums. Small segments can only confirm big effects, which is another reason to start with your largest one.

During the test, only half the segment sees the new page, so you capture about half the incremental revenue until you roll it out.

Which segments usually carry the most upside

Upside is roughly segment size × the gap between what the segment needs and what the default page gives it. Two segments usually score highest.

First-time visitors from paid traffic. On most growing stores this is the biggest segment you pay for. These visitors arrive from an ad with a specific angle, don't know the brand, and often land on a page written for someone who already does. Contentsquare's 2026 benchmark (99 billion sessions, 6,000 sites) puts new visitors at a 1.7% conversion rate. A page that continues the ad's promise and front-loads proof is the classic fix. See how to build a product page for cold traffic and sending Meta ad traffic to a different product page.

Returning visitors and customers. In the same Contentsquare data, returning visitors convert at 2.9%. They already know your pitch. What helps them is a shorter path: restocks, bundles, what's new, the category they browsed last time. Smaller segment, higher value per session. See showing returning customers a different homepage.

Segments that are usually poor first bets: tiny ones (you can't measure them), ones without a distinct need (you can't write a better page for "visitors on Tuesdays"), and ones your email platform already reaches well. For more ideas, see these website personalization examples.

How to measure your own personalization ROI

  1. Pick one segment and one page. Write down the segment's current sessions, conversion rate and RPV for the last 30 to 90 days.
  2. Compute your break-even lift with the formula above, using your margin and costs.
  3. Estimate how long the test needs to run with a sample size calculator. If the answer is longer than you're willing to wait, pick a bigger segment or test a bolder change.
  4. Randomize. Split the segment so a share of visitors (often 50%) sees the personalized page and the rest see the original, at the same time. Before-and-after comparisons don't count: seasonality, promotions and ad changes will swamp the effect.
  5. Measure RPV, not just conversion rate. A page can raise conversion rate by pushing discounted products while lowering revenue.
  6. Decide the end date in advance and don't stop the test the moment it looks good. Early peeks overstate lifts.
  7. Keep a holdout after rollout. Leaving a small share of the segment on the original page lets you confirm the lift holds, and catches a page that stops working.

If you use several personalized experiences, also compare the whole group of visitors in your flows against a holdout that sees none of them. Individual test wins don't always add up, and the store-level number is the one that pays the bills. We cover the mechanics in Shopify rollouts and A/B testing.

Our own numbers, labelled for what they are

Ariven is an on-site personalization app for Shopify: you build alternate versions of your existing pages in the theme editor, set up a flow that says which segment sees which version, and Ariven routes each visitor in that segment to the version you assigned, measuring it against the original page. So we hold ourselves to the standard above.

In our own randomized A/B test on a founder-owned store, three flows split 50/50 against the original lifted measured store sales by 9%, which projects to 18% with all traffic included. Inside the three flows, revenue per visitor rose 160% and conversion rate 98%. Results vary by store. The strongest single flow, a collection landing page for one segment, raised conversion rate 152% and revenue per visitor 205% on that one flow.

Read that the way you'd read any vendor number in the table: it's one store, the founder's own, and we chose to publish it. It is a controlled experiment, which is why we report the store-level 9% alongside the in-flow figures. Your result will depend on your segments, your pages and your traffic, and the only way to know it is to run the same kind of test. If you're comparing tools to run one, see the best Shopify personalization apps.

Sources

Frequently asked questions

What is the average ROI of website personalization?

There is no reliable average. The most quoted figure is McKinsey's: personalization most often drives a 10 to 15% revenue lift, with company-specific lift spanning 5 to 25%. That comes from a consultancy's analysis of companies, not from a controlled experiment, so treat it as a range of what is possible rather than what your store will get.

How much does personalization increase conversion rate?

It depends on what you personalize. In a randomized field experiment with 184,375 users at a large retailer, product recommendations raised final conversion rate by 7.5% (Lee and Hosanagar, Management Science, 2021). Vendor case studies report much larger numbers, but vendors choose which results to publish, and many do not describe a control group.

Are personalization statistics like '80% of consumers want personalization' proof of ROI?

No. Figures such as Epsilon's 80% (1,000 US consumers, 2017) or Twilio Segment's 56% who say they would become repeat buyers (3,001 consumers, 2023) measure what people say in surveys. They show demand for personalization, not how much revenue it adds. Only a randomized test against a control group measures causal lift.

How do I measure personalization ROI on my Shopify store?

Split the visitors in one segment at random: half see the personalized page, half see the original. Compare revenue per visitor between the two groups over a fixed period, multiply the difference by the segment's monthly sessions, then subtract the cost of the tool and your time. Use gross margin, not revenue, if you want profit ROI.

Which visitors give the biggest return from personalization?

Usually large segments with a clear, shared need that the default page ignores: first-time visitors from paid social or a specific ad campaign, and returning visitors, who already convert at about 2.9% versus 1.7% for new visitors across ecommerce (Contentsquare, 2026). Upside scales with the segment's size, so start with your biggest segment, not your most interesting one.