Skip to content
pan paws
All work

01 / 10 · 2026

Presidency University

Presidency wins once students arrive. It was losing where they decide.

Diagnosis delivered
  • Diagnose
  • Funnel analytics
  • SEO
  • AEO/GEO
  • Reputation
  • Strategy
Client
Presidency University, Bengaluru
Role
Growth analytics & digital intelligence: competitive, SEO, AEO/GEO, website, social, admissions funnel, strategy
Engagement
8-week Growth Analytics & Digital Intelligence internship (Jun–Jul 2026)
Year
2026
With
Aditya Kumar, solo

“Think like a consultant, not an intern.” The brief banned generic AI reports, so every finding came from primary analysis: competitor teardowns, live GA4 and CRM data, AI-engine testing and hands-on site audits.

leads → 2,441 enrolments
4,98,603
of leads never verified
75%
ChatGPT prompts that mention PU
0/17
prioritised website fixes
40

01The problem

Presidency University was generating huge lead volume, but almost none of it turned into students. It was also nearly invisible in the places students now decide: Google's high-intent searches, AI assistants and third-party review sites. The university needed to know where students were disappearing, and why, before spending more on ads.

02What I found

Scorecard at the end of the diagnosis · needles show the grade, not a measured score
  • SEO

    Critical

  • AI / GEO

    Critical

  • Funnel

    Mixed

  • Website

    Weak

  • Reputation

    Weak

  • Social

    Developing

01

The funnel: a leaking bucket at the top

  • 01Lead → enrolment was 0.49%. 94.2% of leads never started an application, and about 75% were never verified, so they were unreachable from day one.
  • 02The late funnel was strong: 81% of paid applicants submitted, and 61% of submitters enrolled. The problem wasn't closing students. It was getting them in and getting them to pay.
  • 03Inside the form, most in-progress applications sat at screen one (“Basic & Program Details”). Most applicants never initiated the application-fee payment, and a large group opened the payment step and abandoned it.
  • 04Monthly leads kept rising, but lead → application stayed under 1%. More water into a leaking bucket.
  • 054,98,603 leads → 28,816 started → 4,937 paid → 4,009 submitted → 2,441 enrolled.
  • 063,74,659 leads unverified · 28,963 stuck at form stage 1 · 18,852 never initiated payment · 9,057 abandoned checkout · 1,568 submitted but didn't enrol.

The leaking funnel

One admissions cycle, from first lead to enrolment. Bar lengths are log-scaled: on a straight scale, everything after Leads would be a sliver.

  1. Leads4,98,603Everyone who enquired
  2. Started28,8165.8% of the step before↘ 94% never start an application
  3. Paid4,93717.1% of the step before↘ most never initiate the fee payment
  4. Submitted4,00981.2% of the step before↘ 19% of payers never submit
  5. Enrolled2,44160.9% of the step before↘ 39% of submitters don't enrol
0.49%of leads became students.4,98,603 leads → 2,441 enrolments

02

Tracking: the data broke exactly where students left

  • 01“Apply” handed visitors from presidencyuniversity.in to a separate apply subdomain (the CRM portal), and cross-domain tracking broke at that handoff. The university couldn't see where its biggest leak happened.

03

Which pages actually converted

  • 01Programme-specific pages converted far better than generic ones. The admissions FAQ and the BCA AI/ML page converted well above the homepage.
  • 02High-traffic pages like the BBA programme page, Contact and PhD converted zero.

Note: Small samples, so directional.

04

Search: not ranked where intent lives

  • 01Not ranked for high-intent queries like “BBA colleges in Bangalore”, “best BCA colleges in Bangalore”, “engineering colleges in Bangalore” or “top private university in Karnataka”. Rivals and aggregators owned page one.
  • 02The rankings were brand-based, not intent-based.

05

AI answers: absent where students now ask

  • 01Tested 17 admissions-intent prompts across 4 assistants. PU appeared in 0/17 on ChatGPT, 2/17 on Gemini, 5/17 on Claude and 4/17 on Copilot.
  • 02Entity contamination: one aggregator described PU as “also known as PES University”. Another attributed Presidency College's NIRF #7 to PU. Wrong facts were feeding straight into AI answers.
  • 03Aggregators (Shiksha, Careers360, Collegedunia, CollegeDekho) out-ranked and out-cited the university's own site. Placement figures varied from 55% to 99% across sources, so AI models hedged.
  • 04Strong assets weren't machine-readable on PU's own site: IIRF #33 Pan-India, #1 BBA (Super Excellence), NIRF 2025 Management band.
Illustrative mock-up

Question: Best BBA colleges in Bangalore?

ChatGPT

Answer (illustrative): For a BBA in Bangalore, students often shortlist University A, University B and College C. Compare accreditation, placement records, fees and campus location before you apply.

Presidency University named in 0 of 17 prompts on ChatGPT
The answer text is a generic placeholder, not real AI output. The scores are real: how many of 17 admissions-intent prompts named Presidency University on each assistant, from my test in Jun–Jul 2026.

Aggregators described Presidency University as “also known as PES University”, and credited it with Presidency College's NIRF #7. The fix is to correct the facts first.

Seen on aggregators: “also known as PES University” · “NIRF #7” (Presidency College's ranking, credited to the university).
This isn't a marketing problem. It's a data-integrity problem. Before you can rank in AI, the internet has to agree on who you are.
The reframe

06

Website: the #1 lead source worked against visitors

  • 01The homepage opened on a 13-slide auto-carousel with text baked into images: invisible to search and AI, no reflow on mobile, and no single clear CTA.
  • 02Desktop LCP was about 4.2s (target under 2.5s). Converting images to WebP/AVIF saves roughly 30 MB.
  • 03Key CTAs (“Know More”, “Explore Student Life”, “Hall of Fame”) pointed to #, so clicks led nowhere.
  • 04Programme discovery was buried below a long celebrity “Thought Leaders” wall, and a ~90-programme menu was one flat list.
  • 05Rankings were shown as images, so they couldn't be crawled or cited.

07

Reputation: the story was being told by others

  • 01Only 11 Quora/Reddit threads referenced PU, and about 80% of them were negative or unanswered, with negative sentiment about staff and process.
  • 02Stale 2018–20 complaints still ranked.

08

Social: the strongest selling point was the weakest content

  • 01Scored 186 LinkedIn posts across 13 content pillars, and 48 Instagram posts.
  • 02Placement and careers, the main reason families choose a university, was among the lowest-engagement pillars. Rankings posts drew comments but almost no engagement.
  • 03On LinkedIn engagement per post, PU (67.1) trailed REVA (110.5) and GITAM (91.2), despite having more followers than REVA.

09

Competitors: a better shop window, not a better campus

  • 01Benchmarked Christ, Jain, RVU, Alliance, REVA, Manipal and Symbiosis.
  • 02Their edge was executional: faster pages, richer programme pages, real placement stories, visible fee structures and step-by-step admissions journeys. All of it is buildable.
  • 03When students searched PU's fees, they landed on aggregators instead of the university's site.
Presidency already wins once students arrive. Win the discovery layer (search, AI and reputation) and the funnel takes care of itself.

03What I delivered

  • 017-university competitor teardown, plus a website recommendations backlog with owners and due dates.
  • 02SEO/AEO/GEO audit: keyword-gap analysis across general, Bengaluru and course-specific intent queries, and 4,000+ course-level FAQ questions drafted across 11 schools for AI-answer inclusion.
  • 03A 20-query × 5-engine AI-visibility test panel, an entity-contamination audit and an aggregator-consistency tracker.
  • 04Off-page authority plan: an answer strategy built on 100 curated questions, a review-collection campaign, a faculty thought-leadership programme and a “Myth vs Reality” content series.
  • 0540 prioritised website fixes: 20 UX, 10 conversion and 10 content, plus a mobile, speed and accessibility audit (an auto-rotating carousel with no pause, # links breaking keyboard navigation, weak alt text, low contrast).
  • 06A content framework mapping every format to Reach, Leads or Trust, plus 25 content opportunities.
  • 07Admissions funnel report: the top 5 drop-offs, the top 5 conversion drivers and 10 growth opportunities ranked by impact × effort.
  • 08A 36-month, 100% organic AI-visibility roadmap: zero paid media: entity and data lockdown → correct the record → consolidation → authority.
  • 09A 28-slide strategy deck for university leadership: current state, benchmark, top 20 opportunities, 30-day quick wins, 12-month strategic wins and the 2026–2028 roadmap.

Top fixes recommended

Fix list · ranked by impact × effort
  1. F-01Instant OTP lead verification at captureDo now
  2. F-02Cut the application's first screen to the essentials, with save & continueDo now
  3. F-03WhatsApp/SMS rescue for started-but-unpaid applications, with one-tap paymentDo now
  4. F-04Fix checkout: UPI, EMI, retry on failureDo now
  5. F-05Fix cross-domain GA4 and UTM passthroughDo now
  6. F-06Replace the hero carousel with one fast, real-text hero and a single CTADo now
  7. F-07Show scholarship eligibility before the fee wallNext
  8. F-08Route traffic to programme pagesNext
  9. F-09EducationalOrganization + Course + FAQPage schemaNext
  10. F-10One canonical fact sheet across every aggregatorNext
  11. F-11“PU vs Christ / Jain / Manipal” comparison pagesNext
  12. F-12Lead-source quality routingLater
  13. F-13Faculty research reports for earned mediaLater

04Result

Delivered

  • The diagnosis and the 2026–2028 roadmap, delivered to the university and its digital team.
  • Homepage conversion work is live at presidencyuniversity.in (the next case study). The rest of the roadmap is owned by the university's team.

What I'd measure next

Lead-verification rate, lead → application-start rate, payment-initiation rate, AI query presence on a 40-query panel, and entity accuracy in AI answers.

Method
  • Diagnose
  • Grow
Diagnosis delivered
See the homepage fix that came out of this →

Frames

  • 01Strategy deck: current state
  • 02Funnel report: top 5 drop-offs
  • 03AI-visibility test panel
  • 04Fix list: 40 website fixes
  • 05Roadmap: 2026–2028