Sep 16,2026 
By Mansoor Bhanpurawala TL;DR
An India-based accounting SaaS platform had traffic, but almost none of it was earning new prospects. The site ran on brand searches and one popular free tool, with zero non-branded organic clicks the quarter before this work started. Over the next reporting window, non-branded organic clicks went from 0 to 445 (+4.3% total impression growth, 444,000 → 463,000), and the site picked up 104 organic visitors from AI assistants (ChatGPT, Perplexity, Gemini, Copilot), a GEO result most B2B SaaS SEO case studies can’t show with real numbers. Branded search softened over the same period, an open miss, not a hidden one. Full numbers and sourcing below.
The problem wasn’t “no traffic.” It was the wrong traffic.
An India-based SaaS accounting platform came in with what looked like a surface-level traffic problem.
It wasn’t.
In the quarter before this work began (Jul–Oct 2025), the site pulled 3,980 clicks and 444,000 impressions from Google Search Console.
That’s nothing.
The real problem sat underneath that number.
Almost every one of those clicks came from two sources: people who already knew the brand and searched for it by name, and one high-volume free tool (a quotation maker) that drew traffic but wasn’t built to sell anyone on the actual product.
Strip those two sources out, and what’s left is zero.
Zero non-branded organic clicks.
Zero informational-intent visibility. A prospect searching “free accounting software like tally” or “gst accounting software” had no reason to ever land on this site, because it wasn’t ranking for any of it.
On top of that, nobody had instrumented the signup funnel past a page click. The business believed it had a “no sales” problem.
There was no data to confirm or deny that, because nothing downstream of a click was being measured.
The baseline
Sourced from Google Search Console, 1 Jul–31 Oct 2025, the period immediately before this reporting window.
| Metric | Value |
|---|---|
| Total clicks | 3,980 |
| Total impressions | 444,000 |
| Average CTR | 0.9% |
| Non-branded clicks | 0 |
| Non-branded impressions | 0 |
| Homepage clicks | 1,111 |
Every query that later became a ranking source, “free accounting software like tally,” “gst accounting software,” “lifetime free gst billing software offline,” had zero clicks and zero impressions in this window.
Not low.
Zero.
What we actually did
Built for non-branded, informational-intent search. New and re-optimized content specifically targeting the queries the brand had no visibility on: the “free accounting software,” “GST software,” and comparison-shopping terms a prospect actually types before they know the brand name.
Re-optimized the existing blog library rather than starting from scratch, since several older posts were sitting on page two with fixable title and structure problems.
Built out ICP-specific pages, with mixed results by design. Some landed inside the site’s formal Use Cases navigation (Agriculture, Chartered Accountant, Hospitality). Others, like petrol pump accounting and student-focused content, shipped as blog posts without nav placement. That distinction matters for what follows.
Created two new free tools mid-window: a cost-price calculator and a raw-material calculator, both launched from zero.
Built one live comparison page (vs. Tally). A second (vs. QuickBooks) is live but not yet ranking. A third (vs. Zoho) is planned, not built. It’s not in these numbers.
Build a citation footprint across five platforms: G2, SoftwareAdvice, Capterra.in, SoftwareWorld, and SoftwareSuggest. Worth being precise about what this is and isn’t, more below.
The results
Sourced from Google Search Console and Ahrefs Web Analytics, 1 Nov 2025–28 Feb 2026, compared against the 1 Jul–31 Oct 2025 baseline.
| Metric | Before | After | Change |
|---|---|---|---|
| Non-branded organic clicks | 0 | 445 | new |
| Non-branded impressions | 0 | 52,300 | new |
| Total organic clicks | 3,980 | 4,480 | +12.6% |
| Total impressions | 444,000 | 463,000 | +4.3% |
| Average CTR | 0.9% | 1.0% | +0.1pp |
| AI-assistant visitors | — | 104 | new |
| Registration-page clicks | — | 67 | new |
| Branded clicks (top query) | 675 | 466 | −31.0% |
| Homepage clicks | 1,111 | 818 | −26.4% |
Where it shows up at the page level
On the AI-search number: 104 visitors arrived from ChatGPT, Perplexity, Gemini, and Copilot combined, filtered to organic-only traffic.
ChatGPT alone accounted for 62 of those, with a 2m 41s average session and 62.9% bounce rate, meaningfully better engagement than most of the site’s organic search traffic, for what it’s worth.
Most B2B SaaS SEO case studies talk about GEO in theory. This is one of the few with a sourced, filtered number behind it.
On the directory footprint: five platforms live and indexed is real and worth having. It’s genuine distribution and citation surface area.
What it isn’t yet is a trust signal.
Across all five, there are two distinct human reviewers. One review does duty across three of the five platforms (Capterra explicitly sources its copy from SoftwareAdvice). G2 and SoftwareWorld currently show no reviews at all.
Call this what it is: infrastructure in place, review volume still to build.
On registration-page clicks: 67 unique clicks to the signup page is a real, positive top-of-funnel number. More people reached the door than before. What happens after someone clicks that link isn’t tracked. No signup-completion data exists for this window, so the number in this case study is 67 clicks, not signups, and the completion rate is genuinely unknown rather than estimated.
What I’d do differently
Branded search and homepage traffic both declined over this window: 675 to 466 clicks on the top branded query, 1,111 to 818 on the homepage. The two are connected, fewer people searching the brand name directly means fewer people landing on the homepage as a result. There’s no confirmed cause here, and I’m not going to manufacture one. It’s an open miss, tracked but unexplained, and next on the list.
Signup-completion tracking should have existed from day one. It didn’t. The original brief was “no sales.” But there was never any instrumentation past a page click to actually confirm that, which means the starting problem was assumed, not measured. Fixing that is priority one for the next phase, because right now nobody, including the client, knows the real conversion rate on 67 people clicking through to register.
Two of eight ICP pages underperformed against the rest of the set. They shipped and stayed live without the follow-up pass they needed. That’s a backlog item, not a secret.
FAQs
1. Does non-branded organic traffic going from zero to 445 clicks mean the site had no SEO value before this work?
2. Are the 104 AI-assistant visitors filtered to exclude paid traffic?
3. Does this case study report signup numbers?
4. What makes this a GEO case study and not just an SEO one?
Mansoor Bhanpurawala is an AI SEO Consultant with over 14 years of hands-on experience driving organic search growth for enterprise SaaS brands and high-traffic publications. He holds a Bachelor’s Degree in IT and is certified in Technical SEO and Google Analytics. Throughout his career, Mansoor has audited more than 500 websites, managed content teams, and scaled multiple sites to over 500,000 monthly visitors. His technical insights have been featured in Mid Day, Yahoo, Dailyhunt, and Upcity. You can review his full case studies at digitalmansoor.com or connect with him directly on LinkedIn.