Kunvarani is a mother-and-daughter ethnic wear boutique in Bangalore. Twenty years of buying, a real shop, a small online catalogue. Here is what Sage found in one assessment.
| Segment | Impressions | Clicks | CTR |
|---|---|---|---|
| Traffic she'd written off | 1,312 | 3 | 0.23% |
| Everything else | 402 | 23 | 5.7% |
| Priority range (₹3k+ trousseau kurta sets) | 78 | 0 | 0% |
A quarter of her traffic produced 88% of her clicks. The product she most wants to sell was invisible.
At onboarding she'd told us which searches were worthless to her — wedding-etiquette curiosity, people researching ceremonies rather than shopping. Most SEO engagements never ask that question. It turned out to be the most important input on the account.
The homepage was indexed twice. Google held the http and https versions as separate URLs — 256 impressions and 43 clicks on one, 51 and 1 on the other. The same page competing with itself for her own brand name.
Two near-identical posts were splitting one topic. One on a malformed /blogs/blogs/ path. 457 and 309 impressions, neither ranking better than position 11.
A page-one ranking earning almost nothing. Her best article sits at position 7.2 with 1,717 impressions and a 0.64% CTR. Her homepage, at a comparable position, converts at 16.8% — twenty-six times the click rate. That's a title and snippet problem, and it's cheap to fix.
Her stated reason for excluding that traffic was "zero-click, so zero purchase intent." Sage checked: every excluded query ranked below position 8, where near-zero clicks is expected regardless of intent. The observation was explained by position and said nothing about whether those searchers would buy.
It would not call conversion broken. Zero purchases on fourteen sessions is not a conversion problem — it's not enough data.
It would not dress noise as opportunity. Post-exclusion, the "quick wins" list came to 53 queries totalling 106 impressions. It reported "no material opportunity at current volume" rather than filling a page.
It would not guess at causes it couldn't check. Where a finding needed live server or SERP data the system doesn't collect, it said so.
It also audits its own completed work against the client's stated goals each month, and reports where the two have drifted. The failure mode of an autonomous system isn't doing nothing — it's doing the wrong thing efficiently, for months, while every report looks healthy.
Two of those three are things Sage cannot do itself. It said so, and marked them as needing a human.
Send your store URL. Nothing else — no logins, no access, no call. You'll have your own teardown in 48 hours. If it doesn't tell you something you didn't already know, we're done and you owe nothing.
Send my store URL →Every figure on this page comes from Google Search Console and Google Analytics data for kunvarani.com, 2 May – 31 July 2026, and from Sage's own run records. Published with the client's permission. Kunvarani is Sage's first and founding client.