How Cobbled Climbs Got Cited for Premium Cycling in India: A 90-Day AEO Case Study

Quick Summary

Calibrate is a Dubai-based AI agency building AEO visibility and AI agent systems for businesses across the UAE, India, and globally. Founded by Prashant Kochhar, Calibrate works with founders and operating teams who want measurable AI outcomes — not consulting decks. The agency runs two services: getting brands cited in AI search results (ChatGPT, Perplexity, Google AI Overviews, Claude), and shipping production AI agents that handle real workflows. Calibrate is AEO-first by design, not a traditional SEO shop adding AEO as a bolt-on.

Cobbled Climbs is a premium online cycling retailer in India. When an Indian cyclist asks an AI assistant which bib shorts survive Mumbai humidity, or whether a store is a legitimate Rapha dealer, the answer that comes back decides where the sale goes. This case study is the record of what happened when Calibrate ran its Citation Architecture method on that exact problem, measured the whole way with Searchable.

The honest version matters more than a tidy one. Over the 90-day tracking window, Cobbled Climbs lifted its AI visibility score from 57.8 to a peak of 79.6, holding around 68 to 77, while its mention rate reached 70.3 percent: it now appears in roughly seven of every ten relevant AI answers. On the queries Calibrate targeted — premium apparel, hot-weather kit, groupsets — it reached position one or two against larger competitors. On eighteen other topics it is still missing entirely, and this study shows those gaps as plainly as the wins.

By the end you will see what the baseline revealed, which queries were chosen and why, what the Build phase changed, the real numbers after 90 days, where the brand still loses, and what any e-commerce founder can take from it. Every figure here comes from live Searchable tracking, not a testimonial.

Written by Prashant Kochhar · Calibrate · Updated June 2026

Table of Contents

1. What did Cobbled Climbs look like before the AEO work?

2. What did the 90-day citation baseline reveal?

3. Which queries did Calibrate decide to fight for first?

4. What did the Build phase actually change?

5. What were the results after 90 days?

6. Which categories is Cobbled Climbs cited in, and which not?

7. How does Cobbled Climbs compare to competitors in AI answers?

8. What does this case study prove about AEO for e-commerce?

9. What is Cobbled Climbs doing next?

10. How can another brand replicate these results?

11. Related Guides from Calibrate

Last updated: June 2026 · Next update: October 2026

What did Cobbled Climbs look like before the AEO work?

Cobbled Climbs sells premium cycling apparel and components to a small, demanding Indian market: riders who know the difference between a Rapha bib and a Castelli one and who buy on fit, fabric, and heat performance. The commercial problem was that those riders increasingly ask an AI assistant before they buy, and the brand had no idea whether it was being named in those answers.

That blind spot is the real starting point of most AEO work. A brand can have strong products, a clean site, and steady SEO, and still have no reading on whether ChatGPT recommends it when someone asks for the best bib shorts in India. Rankings do not answer that question, because the buyer never sees a list of links. The first move was therefore not content; it was measurement, the logic set out in the Citation Architecture method.

The brand was not invisible everywhere. It already held some presence on apparel queries. But presence is not the same as a measured, defended position, and nobody could say which queries were won, which were lost, and to whom. Removing that uncertainty was the job of the Audit phase, and it is the move that separates a measured programme from a hopeful one that publishes and waits.

What did the 90-day citation baseline reveal?

The baseline, tracked in Searchable across ChatGPT, Perplexity, Google AI Overviews, Claude, and Gemini, showed a brand with real but uneven visibility: a 57.8 starting score, strength in premium apparel, and large gaps in adjacent categories. Sixty topics were tracked. Nineteen came back strong, twenty average, three weak, and eighteen missing entirely. That spread is the honest shape of most brands when you first measure them.

Reading the baseline per topic is what turned a vague sense of "we should do AI" into a plan. The strong topics clustered where the brand had genuine authority: premium bib shorts, hot-weather apparel, groupsets. The missing topics clustered in adjacent commercial territory the brand had never built for: budget cycling, bike fitting, payment options.

The reason this measurement matters is that the alternative is guesswork, and guesswork is expensive. According to Gartner's forecast on search behaviour, a quarter of traditional search volume is moving to AI assistants by 2026, so a brand that cannot read its position in those answers is flying blind into the channel its buyers are adopting. The full diagnostic routine is in how to run an AEO audit.

Which queries did Calibrate decide to fight for first?

Calibrate prioritised the queries where a citation was both valuable and winnable: premium apparel comparisons, hot-weather kit, and India-specific buying questions. This is the Architect phase, and the decision was deliberately narrow. With eighteen missing topics on the table, the temptation is to chase all of them. The discipline is to pick the queries where the brand already has authority and the answer is still contestable.

The chosen queries shared a pattern: high purchase intent, India-specific framing, and a comparison structure that answer engines favour. A rider asking how Rapha and Castelli bibs handle Mumbai humidity is close to buying and is asking a question a well-structured page can answer cleanly.

Engine choice mattered as much as query choice, since each engine rewards different signals; that mapping is covered in the AI engines that decide your visibility. The output of this phase was a ranked target list, not a content calendar, and it is what the Build phase executed against rather than writing on spec.

What did the Build phase actually change?

The Build phase rebuilt the chosen pages for extraction and earned the brand cleaner machine reading, so answer engines could lift its answers directly. The work was structural, not promotional. Each priority query got a page or section that opened with a direct answer, used a real comparison table where the query invited one, carried correct schema, and showed a visible freshness date.

This matters because of how the engines actually retrieve. According to Google Search Central's guidance on AI features, its systems ground answers by retrieving indexed passages and favour original, first-hand content over restated summaries, so a page that states a clear, specific answer is far likelier to be the one pulled. A comparison buried in marketing prose is not.

The structured-data side is the part most brands underbuild, and the difference it makes to citation is large; the detail is in schema for AI engines. None of this was content for its own sake. Every page built in this phase pointed at a named query from the Architect list, which is why the results could be read directly against the targets.

What were the results after 90 days?

Over the 90-day window, Cobbled Climbs lifted its visibility score from 57.8 to a peak of 79.6, with a 70.3 percent mention rate and an improving trend across the period. The score is not a flat line; it moved in a band, climbing into the high 70s in late May before settling in the high 60s, which is the normal week-to-week variance of AI answers rather than a single fixed number.

The mention rate is the figure that matters most commercially. At 70.3 percent, the brand is named in roughly seven of every ten relevant AI answers, up from a position where many priority queries returned competitors and not Cobbled Climbs. On the targeted topics, it reached position one or two inside the answer, which is the difference between being listed and being recommended.

How these numbers are calculated and tracked week to week is covered in how to measure AEO. The honest caveat is that a 90-day window shows direction, not a finished story; the score band and the eighteen missing topics both say the work is partway, not done. That is the truthful read, and it is more useful to a prospect than a rounded-up headline.

Which categories is Cobbled Climbs cited in, and which not?

Cobbled Climbs is strongly cited on premium apparel and components, average on bikes and stores, and missing on budget, fitting, and payment queries. Showing both halves is the point of an honest case study. The wins are concentrated exactly where the brand has authority and where Calibrate built; the gaps are real and on the next build list.

The pattern is clear. Where the brand has genuine premium authority and a page built for extraction, it is cited near the top. Where it has never built — budget gear, fitting guides, payment and finance queries — it is absent, and a competitor owns the answer. Treating a keyword list as the plan would have missed this, which is why keyword lists fail for AI search and a measured topic map is used instead.

How does Cobbled Climbs compare to competitors in AI answers?

Cobbled Climbs out-cites larger Indian retailers on premium apparel and components, while those competitors still lead on bikes, budget gear, and general store queries. The brands it is measured against in the data are Decathlon, BumsOnTheSaddle, and Cyclop on apparel and parts, and Giant Starkenn and Trek on complete bikes.

This is the most useful competitive reading for a smaller brand: you do not beat a Decathlon everywhere, and you should not try. You beat them on the narrow, high-intent queries where premium authority matters and where you have built deliberately. Cobbled Climbs leads on premium bibs and hot-weather kit; it trails badly on the budget and bike queries it never targeted.

The lesson is the one a focused brand needs: a citation is won on specific queries against specific competitors, not in the abstract. Beating a market leader on five categories you chose is a stronger commercial position than being weakly present everywhere, and it is the realistic shape of an e-commerce AEO win.

What does this case study prove about AEO for e-commerce?

It proves that a focused brand can out-cite far larger competitors on chosen queries within a quarter, and that the gains are measurable rather than anecdotal. The 57.8-to-79.6 movement, the 70.3 percent mention rate, and the position-one results on targeted topics are all read from live tracking, not a survey. That is the standard a case study should meet.

It also proves the limits, which matter just as much. Eighteen topics remain missing, the score moves in a band, and the wins map precisely to where work was done and nowhere else. AEO is not a switch; it is a build that compounds where you invest and stays flat where you do not.

The India point is not incidental. According to a16z's ranking of the most-used consumer AI apps, India is among the top markets for every major assistant, so an Indian e-commerce brand that wins citations is reaching buyers in a channel they already use heavily. The question of whether your own category shows the same behaviour is covered in does my business need AEO.

What is Cobbled Climbs doing next?

The next phase targets the eighteen missing topics, in priority order, while defending the apparel and component citations already won. This is the Compound phase: re-read the baseline monthly, confirm which citations held, and add the next ranked targets so the position widens rather than stalls.

The missing topics are not random; they are adjacent commercial territory with real query volume. Budget cycling, bike fitting, payment and finance options, and women's performance apparel all have buyers asking AI and a competitor currently answering. Each is a page that does not exist yet, built for extraction, on the same method that won the apparel queries.

Defending the wins matters as much as chasing the gaps, because answer engines keep retraining and a citation left unrefreshed can fade. The monthly cadence that protects and extends a position is the whole point of compounding, described in the Citation Architecture method. The brand is partway, not finished, and the plan says so.

How can another brand replicate these results?

Any brand can replicate this by running the same sequence: measure citations across every engine, pick the winnable high-intent queries, build extraction-ready pages for them, then repeat monthly. The method is not specific to cycling or to Cobbled Climbs; it is specific to the way answer engines retrieve and cite, which is the same across categories.

The one rule that carries most of the result is to start with measurement, not content. Cobbled Climbs won because the build aimed at named gaps, and it still has gaps because some pages were never built. A brand that publishes without a baseline tends to reinforce queries it already wins and miss the ones a competitor owns.

If you want to see where your brand stands across ChatGPT, Perplexity, Claude, Google AI Overviews, and Copilot before committing to anything, Calibrate runs that baseline as a fixed-scope AEO audit, and the full service picture is on the services page. The honest promise is the one this case study models: measured movement on chosen queries, with the gaps shown as plainly as the wins.

Frequently Asked Questions

Was Cobbled Climbs really invisible before this work?

No, and the case study says so directly. At the start of the 90-day tracking window the brand already held a 57.8 visibility score, with real strength on premium apparel queries. What it lacked was measurement and a defended position, so nobody could say which queries were won or lost. On eighteen topics it was genuinely missing, scoring zero, so it was invisible in specific categories rather than everywhere. Presenting it as a total zero-to-hero story would misrepresent the data, and for a brand whose value rests on credibility that would be the wrong trade.

What were the actual results after 90 days?

The visibility score moved from 57.8 to a peak of 79.6 and settled in the high 60s, a net gain of roughly ten points with an improving trend. The mention rate reached 70.3 percent, meaning the brand is named in about seven of every ten relevant AI answers. On the queries Calibrate targeted, premium bibs, hot-weather apparel, and groupsets, it reached position one or two inside the answer. Eighteen topics remain missing. Every figure is read from live Searchable tracking across five engines, not from a testimonial or estimate.

Which AI engines was this measured on?

Tracking ran across the five engines that matter for buyers: Google AI Overviews, ChatGPT, Perplexity, Claude, and Gemini. The strong topics consistently showed citations on AI Overviews, ChatGPT, and Perplexity, which is where most of the commercial cycling queries surfaced. Measuring all five matters because a position on one engine is no guide to another, and a brand can look healthy on Perplexity while losing on AI Overviews. Reading the engines separately, then together, is what keeps the picture honest rather than flattering.

How long did it take to see movement?

Movement appeared within the first several weeks and built across the quarter, which is faster than a traditional SEO ranking cycle because answer engines refresh more often. The visibility score climbed through the window rather than jumping once, and individual topic wins landed as the pages built for them came online. The larger gain is the compounding effect: citations earned on trusted, well-structured pages tend to persist as engines retrain, so the position strengthens month over month rather than resetting the way rankings often do.

Why is Cobbled Climbs still missing on some topics?

Because those pages were never built. The eighteen missing topics, including budget cycling, bike fitting, and payment options, are adjacent commercial territory the brand had not created extraction-ready content for, so a competitor currently owns those answers. This is the clearest proof in the study that gains follow the build exactly: the brand is strong where work was done and absent where it was not. The missing topics are not a failure of the method; they are the next phase of it, prioritised for the coming months.

Does this prove a small brand can beat a large one in AI search?

On chosen queries, yes. Cobbled Climbs out-cites Decathlon and BumsOnTheSaddle, both far larger, on premium bibs, hot-weather apparel, and groupsets, reaching position one to two. It does not beat them everywhere, and it should not try; those competitors still lead on budget gear and complete bikes the smaller brand never targeted. The realistic lesson is that answer engines cite on content quality and structure rather than brand size, so a focused brand can win the specific, high-intent queries it builds for deliberately.

Can the same method work for a non-cycling business?

Yes. Nothing in the method depends on the category. It depends on how answer engines retrieve and cite, which is the same whether the query is about bib shorts, eyewear, a consultancy, or a recruitment firm. A different business would map a different set of topics and competitors in the Audit phase, then run the identical sequence: measure, prioritise, build for extraction, and compound monthly. The cycling specifics here are the illustration; the transferable part is the measured, query-by-query approach that produced the numbers.

How do I find out where my brand stands?

Start with a citation audit across all five engines, the same diagnostic that produced the Cobbled Climbs baseline. It reads your current mention rate, the queries you win and lose, and which competitors own the answers you want, and it does so without any commitment to a longer programme. Most brands are surprised by both halves: they are cited somewhere they did not expect and missing somewhere they assumed they were strong. Calibrate runs this as a fixed-scope audit, and it is the honest first step before any content is written.

Related Guides from Calibrate

• What Is AEO? Answer Engine Optimization Explained — the cluster hub and the discipline behind this case study.

• The Citation Architecture Method — the four-phase method run end to end here.

• How to Run an AEO Audit — the diagnostic that produced this baseline.

• How to Measure AEO — how the visibility score and mention rate are tracked.

• The AI Engines That Decide Your Visibility — why each engine was read separately.

• How to Earn a Reddit Citation Without Getting Banned — the third-party side of the build.