Product Analytics Case Study

The Fulfillment Problem

What 274,902 Public Reviews Reveal About Trust in Home-Service Marketplaces

Determine whether fulfillment failures visible in individual customer experiences also appear at scale in public review data across home-service marketplaces.

By Shreyash Dubey · A product analytics case study based on public Google Play Store reviews from Pronto, Snabbit, and Urban Company.

Objective

Test whether fulfillment and trust issues observed in usage are reflected in large-scale public review data.

Dataset

274,902 Google Play reviews across Pronto, Snabbit, and Urban Company (India).

Methodology

Python scraping, rule-based keyword classification, sentiment tagging, and cross-platform aggregation.

Top Findings

  • 274,902 reviews analysed across three platforms
  • Nearly 50% of Pronto negative reviews referenced fulfillment failures
  • Reviews mentioning no-shows were 4.46× more likely to become 1-star reviews

49.7%

of negative reviews referenced fulfillment failures

Based on 9,571 negative Pronto reviews (Jan–Jun 2026)

Section

Key Metrics

What are the headline numbers from the analysis?

Dataset Metrics

Reviews Analyzed

274,902

Total public reviews across three platforms

Pronto Reviews

47,434

Full Pronto review corpus on Google Play

Operational Metrics

2026 Negative Reviews

9,571

Pronto 1–3 star reviews in 2026

No-Show Complaints

44.9%

Primary theme among Pronto negative reviews

No-Show + Support Overlap

40.9%

No-show complaints that also mention support

Risk Metrics

Fulfillment Risk

49.7%

Share of negative reviews citing fulfillment failures

Relative Risk

4.46×

Higher likelihood of a 1-star rating after no-show

Severity Score

67.1

Composite severity index for fulfillment complaints

Section

Dataset Overview

What data was collected and from which platforms?

PlatformReviews
Pronto47,434
Snabbit9,908
Urban Company217,560
Total274,902

Public Google Play Store reviews collected using Python and google-play-scraper.

Section

Methodology

How was public review data transformed into structured insights?

A reproducible pipeline from raw Google Play reviews to classified complaint themes and cross-platform benchmarks.

Google Play Reviews

Python Scraper

Cleaning

Keyword Classification

Sentiment Analysis

Aggregation

Visualisation

Business Insights

Python
Pandas
Plotly / Recharts
google-play-scraper
NLP
Manual keyword validation

Section

Evidence Explorer

What do individual reviews look like behind the aggregate metrics?

Filter anonymised public reviews by theme, rating, or keyword to inspect the underlying customer language.

15 of 15 reviews

R001
No-Show
+ Payment
2026-06-17

fraud people 0 servive suport when you make the booking on app there is no instration after payment is maid all the glich starts worst company not giving service even not ready to refund even there is no one at custome support only chat boat

R002
No-Show
2026-06-17

booking karke payment le lete hai aur jab time hota hai to delay ka msg pop-up hota hai aur koi update nahi milta

R003
No-Show
+ Payment
2026-06-17

very poor experience continuously cancelled for3 days because they don't have staff and they don't refund money BEWARE OF THE TRAP

R004
No-Show
+ Support
2026-06-16

They are literally now scamming people. I booked service and no one was assigned. i then rescheduled, for which a lady named Negma was assigned. they started the service from backend and never came to my location. there is no one you can call . no support or customer service exec...

R005
No-Show
+ Incomplete
2026-06-16

Literally you guy's are very bad service provider, you can't do single msg also what is going wrong there, my slot time was 6:30 pm and now time is 07:03 you guy's still not assigned any buddy. very bad service.

R006
Support
2026-06-16

this is a pathetic app and the services they take the booking and payment but no services they are giving to the customer at assign time. there is no customer support giving assistance to their customers. chat support is not working

R007
Support
2026-06-16

Booked service before 3 days. But no service individual was available at the scheduled time. The customer care executive kept insisting on rescheduling the service after 3 days.

R008
Support
2026-06-16

terrible experience. you write there executive joining the chat, but never did. you can cancel our booking anytime but I can't reschedule or cancel it. bad bad bad support

R009
Incomplete
2026-06-16

Pronto is not reliable even if you book service in advance. The professionals are not trained, they are really slow and would waste time when nobody is monitoring them.

R010
Incomplete
+ Support
2026-06-14

Pathetic service and customer care is like horrible. Seriously go with Snabbit or Urban more reliable ones , don't go with this app because this is cheaper.

R011
Incomplete
+ Payment
2026-06-13

pathetic service. left after taking OTP. misbehaved. broken utensils at my place. completely unsafe experience and pronto didn't provide refund even nor took any actions against her

R012
Payment
2026-06-17

advance booking advance payment kar ke bhi nahi aate aur busy shedule likh kar cancel kar deta hai jab kam ka load hai customer ko weight kyu karvate hai

R013
Payment
2026-06-15

There is a big scam going on with this app....these are charging for services and then cancel themselves and do not refund money. These are looting customers.

R014
No-Show
2026-06-10

Service was delayed by over an hour. Professional arrived late and rushed through the work. Expected better for a prepaid booking.

R015
Other
2026-06-08

Average experience. Sometimes good sometimes bad. Depends on which professional you get assigned.

Section

Representative Customer Reviews

How do real customers describe fulfillment failures?

Anonymised excerpts from the tagged review corpus, with complaint categories and highlighted keywords.

No-Show
+ Support
2026-06-16

They are literally now scamming people. I booked service and no one was assigned. I then rescheduled, for which a lady named Negma was assigned. They started the service from backend and never came to my location. There is no one you can call. No support or customer service exec...

assignedsupportcall
No-Show
2026-06-16

Literally you guy's are very bad service provider, you can't do single msg also what is going wrong there, my slot time was 6:30 pm and now time is 07:03 you guy's still not assigned any buddy. very bad service.

assignedslottime
Support
2026-06-16

Booked service before 3 days. But no service individual was available at the scheduled time. The customer care executive kept insisting on rescheduling the service after 3 days.

customer carescheduledrescheduling
Incomplete
2026-06-16

Pronto is not reliable even if you book service in advance. The professionals are not trained, they are really slow and would waste time when nobody is monitoring them. After their service they would also pressurize you to give a 5 star rating.

trainedslowreliable
Payment
2026-06-17

Advance booking advance payment kar ke bhi nahi aate aur busy schedule likh kar cancel kar deta hai jab kam ka load hai. Time waste app.

advancepaymentcancel

Section

Complaint Distribution

What are customers complaining about when they leave negative reviews?

Pronto Complaint Themes (2026 Negative Reviews)

What are customers complaining about when they leave negative reviews?

Nearly half of all negative reviews referenced fulfillment failures.

Interpretation: No-show and scheduling issues dominate the complaint landscape — not pricing or app UX alone.

Takeaway: Fulfillment is the primary driver of negative sentiment in this sample.

Section

Supporting Analysis

What additional patterns reinforce the primary fulfillment finding?

Rating polarisation, temporal stability, language patterns, and cross-platform comparisons.

Review Rating Distribution (Pronto)

How polarised is customer sentiment in the review corpus?

71% of reviews are 5-star; 21% are 1-star — middle ratings are thin.

Interpretation: Customers tend toward strong opinions. Negative experiences cluster at 1-star, amplifying reputational impact.

Takeaway: The rating distribution is bimodal — failures convert to the harshest ratings.

Fulfillment Complaint Trend (2026)

Has fulfillment risk changed over the analysis period?

Fulfillment complaint share remained near 50% throughout Jan–Jun 2026.

Interpretation: The issue appears persistent rather than episodic — not confined to a single release or month.

Takeaway: Fulfillment risk is stable and structural in the 2026 negative review sample.

Top Terms in No-Show Reviews

What language do customers use when describing fulfillment failures?

Top terms — service, time, booking, support — reflect operational friction, not product features.

Interpretation: Review language centres on execution and recovery, aligning with classified complaint themes.

Takeaway: Text patterns corroborate the keyword classification approach.

No-Show + Support Overlap by Platform

Do fulfillment failures also trigger support friction in the same review?

40.9% of Pronto no-show reviews also mention support — nearly 2× competitors.

Interpretation: Fulfillment failures may cascade into recovery interactions, compounding dissatisfaction.

Takeaway: Support co-occurrence is highest on Pronto, suggesting a fulfillment-to-support chain.

Negative vs Positive Review Share

What share of each platform's corpus is negative?

Pronto and Snabbit show ~20% negative share in their respective corpora; Urban Company's volume skews positive.

Interpretation: Cross-platform star averages differ — comparisons should focus on complaint structure, not raw star averages.

Takeaway: Normalise by complaint theme, not headline rating alone.

Complaint Severity Distribution

How severe are the worst complaint combinations?

40.9% of no-show reviews also reference support — classified as high severity.

Interpretation: The most damaging reviews combine operational failure with recovery friction.

Takeaway: Compound complaints drive the highest reputational damage.

Section

Fulfillment Risk Benchmark

Is Pronto's fulfillment risk an outlier compared to competitors?

Fulfillment Risk Across Platforms

Is Pronto's fulfillment risk an outlier compared to competitors?

Fulfillment-related complaints emerged as the dominant source of dissatisfaction across all three platforms.

Interpretation: Pronto shows the highest fulfillment risk share, but the pattern is category-wide — not isolated to one app.

Takeaway: Fulfillment risk is elevated for Pronto and present across the category.

Section

Relative Risk Analysis

How strongly do no-show complaints correlate with 1-star ratings?

Relative Risk of Receiving a 1-Star Rating

How strongly do no-show complaints correlate with 1-star ratings?

A Pronto review mentioning a no-show was approximately 4.5× more likely to become a 1-star review than the average review.

Interpretation: Fulfillment failures are not mild inconveniences — they strongly predict the most damaging ratings.

Takeaway: No-show complaints carry disproportionate reputational risk.

Section

Complaint Concentration

Are complaints concentrated in a few themes, or spread evenly?

Pronto

0.327

Herfindahl Index (HHI)

Snabbit

0.332

Herfindahl Index (HHI)

Urban Company

0.343

Herfindahl Index (HHI)

Insight: Complaint concentration patterns were remarkably similar across platforms.

Interpretation: Similar HHI scores suggest common marketplace challenges rather than platform-specific anomalies.

Takeaway: The complaint structure is structurally comparable across competitors.

Section

Complaint Journey

What is the downstream cost chain when fulfillment fails?

Booking

Customer confirms slot and payment

Professional No-show

Assigned partner fails to arrive or start

Customer Contacts Support

User seeks resolution via chat or call

Refund / Delay

Resolution is delayed, partial, or denied

Negative Review

Customer publishes a 1–3 star review

Trust Loss

Brand credibility erodes publicly

Retention Risk

Repeat booking probability declines

Review data suggests fulfillment failures generate downstream support demand and may impact long-term trust.

Each step amplifies the previous — the cost extends beyond the missed appointment.

Takeaway: Fulfillment is an upstream lever with downstream retention consequences.

Section

Key Findings

What are the executive-level conclusions supported by the evidence?

Finding

Nearly half of all negative reviews referenced fulfillment failures.

Evidence

49.7%

Business Interpretation

Fulfillment emerged as the dominant operational complaint in the analysed dataset.

Finding

274,902 public reviews were analysed across three platforms.

Evidence

274,902

Business Interpretation

Sample size supports statistical comparison across Pronto, Snabbit, and Urban Company.

Finding

40.9% of no-show complaints also referenced support.

Evidence

40.9%

Business Interpretation

Fulfillment failures frequently cascade into support interactions within the same review.

Finding

No-show complaints were associated with a 4.46× higher likelihood of a 1-star rating.

Evidence

4.46×

Business Interpretation

The most severe ratings disproportionately follow operational failures, not minor issues.

Finding

Similar complaint structures appeared across competitors.

Evidence

HHI 0.33

Business Interpretation

Category-wide patterns suggest shared marketplace dynamics rather than isolated platform issues.

Section

Questions Suggested by the Data

What operational areas warrant further investigation with internal data?

Analytical hypotheses — not product recommendations. Each area lists metrics that could validate or refute the pattern.

Faster Reassignment

Does assignment latency predict no-show rates?

Assignment latency
Completion rate
No-show rate

Recovery Speed

Does resolution time affect rebooking after a failure?

Resolution time
Retention
Rebooking rate

Reliability Scoring

Can partner reliability scores predict fulfillment complaints?

On-time arrival
Completion rate
Customer rating

Completion Verification

Do completion disputes correlate with incomplete-work tags?

Completion disputes
Support escalations
Refund requests

Section

Scope & Limitations

What are the boundaries of this analysis?

  • Public Google Play reviews only — App Store data was unavailable via public APIs
  • No access to internal company operational data (assignments, SLAs, support tickets)
  • Observational analysis — correlation does not imply causation
  • Keyword-based categorisation with manual validation; some misclassification is expected
  • Insights should be validated against internal operational metrics before product decisions
  • Review text reflects self-selected, vocal customers — may over-represent extreme experiences

Section

Technical Appendix

Tools, technologies, and links to the full analysis artifacts.

Technologies Used

Python
Pandas
NumPy
Google Play Scraper
NLP
React
Next.js
Tailwind
Recharts
Vercel

Built by Shreyash Dubey