Skip to content
Product
PlatformPricingDimensionsIntegrationsSecurityCompare
Solutions · by use case
Customer, guest and fan experienceProduct feedback and prioritizationRevenue, churn and retentionAgenciesContact center performanceAI agent and chatbot analyticsMulti-location performanceReviews and competitive benchmarking
Solutions · by industry
Hotels and hospitalityRestaurants and food serviceFitness and leisureRetail and e-commerceSports and live entertainmentInsuranceBanking and financial servicesSoftware and SaaSAutomotiveConsumer products and CPGContact centers and BPOsPublic sector and nonprofitReal estateGaming and betting
Learn
Artifacts GalleryCustomer storiesResearchBlogThought leadershipLessons Log in Start now
Blog

How Do PE-Owned Multi-Unit Brands Measure Customer Experience Across a Portfolio?

Brand NPS scores do not add up across a portfolio. Here is the shared measure that does, how to compare a gym with a hotel fairly, and what the owner's report should show.

Private equity owners of multi-unit brands measure customer experience across a portfolio by reading every unit's reviews, surveys and support contacts with one shared set of fields, turning that into one measure every brand has (the share of customers who had a bad visit), and comparing each unit with competitors of its own category in its own market rather than with its sister brands. The owner then gets a single view with a gap and a cause for every unit, and each brand keeps the fields its own operation needs. Dimension Labs is built for that job: it reads every channel of every brand, adds Dimensions written for each business, joins the results to unit sales and labor data, and builds the portfolio deck the operating partners read.

This post covers why brand scores do not add up across a portfolio, the measure that does, how to compare a gym with a hotel fairly, and what the owner's report should hold. The examples use three of our public-review analyses and a labelled synthetic portfolio. If your portfolio is hotels under different management companies, comparing guest reviews across operators covers that narrower case; for locations of a single brand, see how to compare NPS across locations fairly.

Why don't brand NPS and star ratings add up across a portfolio?

Because each brand measures something different. A fitness brand may run a member NPS survey, a quick-service brand a receipt survey with a five-point satisfaction scale, and a hotel group a post-stay survey plus Tripadvisor. The questions, the scales, the response rates and the people who answer all differ, so an average of the three is not a portfolio score.

Categories also start from different places. Customers of a gym and customers of a hotel complain at different base rates, so a brand-level score mostly tells you which category it is in. And a brand average hides the units. Economists studying restaurant hygiene found that franchised units of a chain tend to free-ride on the chain's reputation, a unit-level problem that a brand-level number does not show.

Ratings also work differently for chains. Michael Luca's Harvard Business School study of Yelp found that a one-star increase led to a 5 to 9 percent increase in revenue, that the effect was driven by independent restaurants, and that ratings did not affect restaurants with a chain affiliation. We put that study next to three others in what a drop in star rating costs. For a portfolio owner the useful question is less "what is our rating" and more "which units are letting customers down, and why".

What measure works across every brand in a portfolio?

The share of customers who report a bad experience, read from what they wrote, with the same core fields for every brand. Every brand gets public reviews, so every unit has the raw material even where no survey runs. A one- or two-star review is the simplest version; a field such as "had a bad visit" read from survey comments and support contacts extends it to channels without stars.

Split the fields in two. Core fields are the same everywhere: staff behavior, cleanliness, wait, accuracy of the order or booking, billing and fees, and whether the customer says they will not return. Brand fields cover what only that operation has: equipment and cancellations at a gym, drive-thru and order accuracy at a quick-service brand, room condition and check-in at a hotel. Core fields let the owner compare; brand fields let each brand act.

Our public analyses show why the unit, not the brand, is where the work is. In each one the spread between units of the same brand was wide, and the failure behind it was different in each category.

Dimension Labs analysis

Inside one brand the spread between units is wide, and the failure behind it differs by category

Three of our public-review analyses, each read with one fixed set of fields per review. Best and worst unit by share of unhappy reviews, and the strongest driver each analysis found.

Public analysisBest to worst unitStrongest driver
Orlando hotels
45 hotels, 9,446 reviews
7.89% to 41.18% poor staysRude or dishonest staff, a wrong charge: +44 pts each (Exhibit D)
Chicago gyms
33 clubs, 9,186 reviews
3% one-star (outer suburban) to 34% (urban), by club typeRude or absent front desk: +36 pts (26 to 52)
Atlanta stores
24 stores, 2,354 reviews
14.7% to 32.8% unhappy (Exhibit 3.1)No help on the floor: impact 9,849 vs 2,278 next (Exhibit 2.1)
From the public pages for the Orlando hotel, Chicago gym and Atlanta store analyses. All public Google reviews. Points are percentage points on the chance of a one- or two-star review; the store figure is an impact score (how often times how badly), so compare within a row, not across rows.

The Atlanta stores make the point most cleanly, because all 24 share a brand, a city and a housing market and still run from 14.7% to 32.8% unhappy reviews, the top figure on 67 reviews and flagged in the report to read with caution. A portfolio report that shows only the brand average would put all of them at 25.7%.

From the report: Atlanta home-improvement store analysis

Exhibit 3.1 of the Atlanta store analysis: all 24 stores ranked by share of unhappy reviews, from 32.8% to 14.7%, against a 25.7% average, with stores under about 70 reviews flagged
Exhibit 3.1 of the published report. Same brand, same city, same housing market. Open the full report

The way each analysis tagged its reviews, with fields fixed before reading and a margin on every unit, is the same method you would run across brands; how to analyze Google reviews across many locations walks through it step by step.

How do you compare a gym brand with a hotel brand fairly?

Do not compare them with each other. Compare each unit with competitor locations of its own category in its own market, and report the gap. A gym with 24% poor reviews in a market where rival gyms sit at 30% is doing well; a hotel at 19% where rival hotels sit at 11% is not. The gap puts every unit of every brand on one scale that means "better or worse than the customer's alternative".

The worked example below uses invented data for nine units across three brands. Ranked on the raw share of poor reviews, the gyms fill three of the top five places, because gym reviews run more negative everywhere. Ranked on the gap to category peers, Gym Unit 4 moves from second-worst to best, 24% against a peer rate of 30%, and Hotel Property B moves to the top of the list, 8 points worse than the hotels around it.

Worked example (synthetic data)

Ranked against its own category's peers, the gym that looked second-worst comes out best

Nine units from three brands in a made-up portfolio. Left: share of reviews rated one or two stars. Right: that share minus the share for competitor locations of the same category in the same market. Higher is worse on both sides.

By raw share of poor reviewsBy gap to category peersGym, Unit 931% poorGym, Unit 424% poorQuick-service, Unit 1221% poorHotel, Property B19% poorGym, Unit 218% poorQuick-service, Unit 717% poorQuick-service, Unit 315% poorHotel, Property A12% poorHotel, Property C8% poorHotel, Property B+8 ptsQuick-service, Unit 12+7 ptsQuick-service, Unit 7+5 ptsHotel, Property A+2 ptsGym, Unit 9+1 ptsQuick-service, Unit 3+1 ptsHotel, Property C-2 ptsGym, Unit 2-4 ptsGym, Unit 4-6 ptsBy raw share ofpoor reviewsBy gap tocategory peersGym, Unit 931% poorGym, Unit 424% poorQuick-service,Unit 1221% poorHotel, PropertyB19% poorGym, Unit 218% poorQuick-service,Unit 717% poorQuick-service,Unit 315% poorHotel, PropertyA12% poorHotel, PropertyC8% poorHotel, PropertyB+8 ptsQuick-service,Unit 12+7 ptsQuick-service,Unit 7+5 ptsHotel, PropertyA+2 ptsGym, Unit 9+1 ptsQuick-service,Unit 3+1 ptsHotel, PropertyC-2 ptsGym, Unit 2-4 ptsGym, Unit 4-6 pts
Synthetic data, invented for illustration (portfolio_example.py). Margins of error on each unit's own rate run from 1.7 to 4.5 points at these review counts.

Three details keep the comparison honest:

  • Get the peer rate the same way as your own. Read competitors' public reviews with the same fields and the same time window, so the gap is not an artifact of different tagging. Review benchmarking covers reading rivals side by side.

  • Put a range on every gap. At a few hundred reviews a unit's own rate is uncertain by several points; in the example, Quick-service Unit 7 is 5 points behind its peers on 300 reviews, with a margin of 4.3 points on its own rate alone. Flag thin units rather than ranking them.

  • Use a recent window. The last 12 months reflect the current operator and the current playbook; an all-time rating reflects the old ones.

Which channels should a portfolio CX read include?

Public reviews are where to start, because every unit has them and the owner does not need anyone's permission to read them. That is also how many deal teams first read a target; customer due diligence from public reviews shows the pre-close version. After close, add what each brand already collects: survey comments, support tickets, chats, call notes and cancellation reasons. Each channel goes through the same core fields, so a brand with a strong survey program and a brand with none are still compared on the same terms. Combining reviews, surveys, tickets and calls covers the file layout.

Then join the read to what the owner already tracks by unit: sales, traffic, labor hours and staffing, opening date, remodel date and manager tenure. That join is what turns "Unit 12 has more complaints about waits" into "Unit 12's wait complaints rose when its labor hours fell". How to combine reviews with POS sales by location shows the join on store IDs.

What should the owner's portfolio report show?

  • One row per unit across every brand: the poor-experience rate, the peer rate, the gap with its range, and the review count.

  • For each unit behind its peers, the two or three core fields that explain most of the gap, so the conversation with the brand team is about a specific failure.

  • The change since last quarter for each unit, so improvement is visible even where the gap is still negative.

  • A brand page for each brand team with its own fields, and a one-page summary for the investment committee.

  • The same template every month or quarter, so the numbers can be compared over time.

In Dimension Labs the agent builds this as a deck, a PDF or a dashboard from the enriched data, reruns it on a schedule you set, saves finished reports to the Final Reports folder, and shares them by link.

Why do portfolio owners run the cross-brand read in Dimension Labs?

Dimension Labs is an omni-channel analytics agent for customer data. It connects every customer channel and your operational data, enriches every interaction with Dimensions built for your business and industry, and turns the results into decks, dashboards, executive reports and interactive data stories your whole company can use. Plans start at $299 a month with no demo.

  • Fully omni-channel: public reviews for every unit, then each brand's surveys, tickets, chats and calls, all read with the same core fields, so no brand looks better because it uses a different survey tool.

  • Easy integration across customer channels and operational data: a unit table with brand, category, market and opening date joins on your unit ID, and sales and labor join by unit and period. The join to revenue data and effects with confidence intervals are on Prove ($2,999 a month) and up.

  • Effortless enrichment of every interaction: shared core Dimensions for the owner and brand-specific Dimensions for each brand team, so the operating partner and the gym's head of operations read the same data with the fields each needs. Custom Dimensions start on Discover ($999 a month).

  • A new visual analytics layer: the unit league table against peers, a page per brand and the committee summary, built in the platform and shared across the portfolio.

Travelers, a Fortune 100 insurer, put every customer interaction across its lines of business onto one schema so it could compare performance across business units and regions, and reports 600% ROI within 18 months and $3.13M in total savings. A portfolio of brands is the same problem with units in place of business lines. See Dimension Labs for multi-location operators.

FAQ

What is the best way to measure customer experience across a PE portfolio of brands?

For an owner who needs every brand's reviews, surveys and support contacts read with the same fields, compared with category peers and joined to unit data, Dimension Labs. Whatever you use, pick one measure every brand has (the share of bad experiences), compare each unit with competitors in its own category and market, and show a range on every gap.

Can I average NPS across brands to get a portfolio score?

Not meaningfully. Each brand's survey has different questions, response rates and respondents, and categories differ in their base rates. Report each unit's gap to its category peers instead, and roll that up if the committee wants one number.

How many reviews does a unit need before it can be ranked?

Enough that its range is narrower than the gaps you care about. Our Atlanta store analysis flagged stores under about 70 reviews to read with caution. Pool small units over a longer window and show them as "not enough reviews yet".

Do the brands have to change their survey tools?

No. Each brand keeps its tools; the raw responses, reviews and tickets are read with the same core fields, which is what makes them comparable.

To see your own portfolio on one page, export each brand's reviews with a unit ID and date, add a unit list with brand and category, and start with one file. Every plan from Explore to Scale has a 14-day free trial.

Dimension Labs/Start with one file From $299 a month

Stop reporting. Start understanding.

Start with one file. Ask why a number moved. Plans from $299 a month, no demo and no contract on monthly plans.