Sahasika
Product pages that were slow, hard to shop on a phone, and invisible to Google Shopping.
Fashion — men's ethnic and embroidered wear
This story describes what was shipped, not a measured lift. No conversion, traffic or revenue figures appear here — see the note on numbers on the case studies hub.
The problem
Embroidered suits and silk-blend kurtas — a catalogue where the product image is the sales pitch, on product pages that made the images hard to reach.
The image gallery did not support swipe, which on a phone is not a missing nicety but a dead end: the gesture every shopper tries first did nothing. The sticky Add-to-Cart button had regressions, so the one control that has to stay reachable did not. Tap targets sat tight in the thumb zone.
Underneath the friction, the catalogue had a discoverability problem: missing Product structured data, thin titles and descriptions, and around 50 products missing barcodes — which is to say, invisible to Google Shopping regardless of how good the product was. And sessions were leaking with nothing in place to recover them.
What we fixed
Applied to the catalogue and shipped live:
- Around 13 page titles and 8 meta descriptions
- Image alt text
- Barcodes, closing the Google Shopping gap
On-page friction: CTA visibility improvements and a form-field fix applied; the mobile carousel-swipe and sticky Add-to-Cart failures surfaced with fixes attached.
Retention layer built: exit-intent rescue, cart rescue (free-shipping bar plus cart reminder), email-capture popups, a lead-capture quiz, returning-visitor offers, and seasonal event campaigns.
Agent-assisted merchandising: product updates executed, AI-generated product images, a price rule, and discounts — all through DynoAgent on approval.
One finding worth the install on its own
While working the data, the agent surfaced a channel observation the store had not acted on: Google.com was converting at roughly 2.6× the store average. That is not a result DynoWeb produced — it is a fact about where the store's existing budget was already working hardest, which is the kind of thing that stays buried in a blended acquisition report.
How we fixed it
Five CRO Reports mapped the ground
Funnel, frustration signals and mobile friction, ranked into a specific plan rather than a list of observations.
SEO Autopilot applied the catalogue fixes
Titles, descriptions, alt text and barcodes, each written through the Admin API and read back to confirm it landed.
SmartNudge built the recovery layer
Exit-intent through to seasonal campaigns, assembled from templates rather than built from scratch.
DynoAgent did the Shopify grunt work
Product edits, image generation, and the discount and price-rule creation — executed after approval.
The end result
Product pages became discoverable to Google Shopping and materially less frustrating on a phone. A full retention layer was configured. Product content and imagery were refreshed through the agent. And the store came away with a data-backed pointer on where its acquisition budget was already working.
Status check before quoting this
Several of the interventions above are built as drafts in the editor rather than published. Anything claiming a nudge is live and earning needs the store's current publish status checked first.
How DynoWeb helped
One app playing CRO analyst, SEO team and retention marketer.
A small fashion team has none of those three in-house and cannot justify hiring any of them. The work above is not clever — it is a lot of separate, ordinary, skilled tasks that nobody on the team had time to sequence. DynoWeb found the leaks, applied the fixes it could apply safely, built the recovery nudges, and had the agent do the repetitive Shopify editing.
Features used
CRO Report
Five audits mapping funnel, frustration and mobile friction into a ranked plan.
SEO Autopilot
Applied the titles, descriptions, alt text and barcodes that Google Shopping needs.
SmartNudge
Exit-intent, cart rescue, email capture, quiz and seasonal campaigns.
DynoAgent
Product edits, generated imagery, discounts and a price rule — on approval.