Written by Ram Kr Shukla. Part of the Shopify SEO library.

Here’s a pattern that’s easy to miss in an Indian store’s Search Console. A shopper wants to know what a lip tint is. She doesn’t type what is a lip tint. She types lip tint kya hota hai. Your English explainer ranks for it, sits around position 6 or 7, and gets scrolled past.

The ranking isn’t the problem. The snippet is.

On one Indian cosmetics store I audited (a D2C beauty brand on Shopify), the explainer posts were catching 207 Hindi or Hinglish query rows over a trailing 28 days, India only, brand terms removed. Those rows added up to 3,735 impressions and 6 clicks, at an average position of 7.7. That’s page one, and almost nobody clicking.

This guide is the fix I’d make on any Indian D2C store, in the order I’d make it. It sits in the Shopify SEO hub next to five other fixes from the same audit. You’ll get a regex for Search Console that actually separates Hinglish from English, the one-line page change, the narrow cases where a title or meta edit is worth it, and the reasons a Hindi site is usually the wrong answer. Every Google behaviour I state links to Google’s own documentation. Where Google hasn’t said something, I’ll say so.

The Gap: Ranked on Page One, Almost Never Clicked

These were the biggest rows in that store’s Hinglish segment. Every one of them is a generic category question, and every one landed on a what is explainer post that answered it well, in English.

You don’t need an industry CTR curve to see the problem. 6 clicks from 3,735 impressions is a 0.16% click-through rate. I’m not going to quote you a benchmark for position 7, because the published curves vary a lot by query type and none of them are about Hinglish. 3,735 chances, 6 takers. That’s enough.

When I pulled the same property’s query list again while writing this, three more things stood out.

None of this is a reason to panic about rankings. It’s a reason to look at what the searcher sees. If your blog is doing the job described in the Shopify blog engine guide, it’s already catching the informational demand. This is about converting the part of it that arrives in Hinglish.

Why the Snippet Loses the Click, and What Google Has Actually Said

Put yourself in the searcher’s seat. She typed a question in Hinglish. The result at position 6 shows a title and two lines of polished English that never echo her words. The result above it might look closer to what she asked. Or she taps the Hindi results chip. Either way, your page didn’t look like the answer.

Here’s what Google has documented that’s relevant, and nothing more.

What Google hasn’t published is how it matches a Romanised Hindi query to an English page, or any recommended way to target Hinglish. So treat everything below as practice checked against real Search Console data. It isn’t a Google rule, and I won’t dress it up as one.

How to Find Hinglish Queries in Search Console

Search Console’s Performance report does support regex. Google’s help page on advanced filtering spells out the details: choose Queries or Pages, pick Custom (regex), and it uses RE2 syntax. Matching is partial by default, so a pattern matches anywhere in the query unless you anchor it with ^ or $. It isn’t case-sensitive by default; prepend (?-i) if you ever need it to be.

The starter regex

Paste this into a Query filter set to Custom (regex), Matches regex:

\b(kya|kaise|kaisa|kaisi|kyun?|kaun|kaunsa|konsa|kitne|hota|hoti|hote|hai|hain|matlab|liye|wala|wali|wale|chahiye|sabse|mein|antar)\b|hindi

The \b on each side is a word boundary, which RE2 supports as an ASCII word boundary. Don’t skip it. On that store, a plain Queries containing hai filter pulled in hair queries and a person’s name. Containing ke matched every query with the word makeup in it. The boundaries stop a token from matching inside a longer English word.

What the tokens cover:

I left some common words out on purpose. me matches near me, which is in plenty of English queries. ka, ke, ki and se are real Hinglish connectors, but they’re short enough to collide with abbreviations and shade codes. And h (as in kya hota h) is too short to trust. The question word usually catches those rows anyway.

On that store’s real query rows, the starter pattern caught the kya hota hai and meaning in hindi families, the ke liye questions and the kaunsa questions. It didn’t touch the hair and makeup queries that broke the plain contains filter. Paste it into your own property once and skim the first 50 rows before you trust it, though. Every catalogue has its own collisions.

Three extra patterns, run separately

Run these as separate filters so the noise doesn’t pollute the clean list.

The wider net adds the short connectors. Expect some English noise and read the rows:

\b(ka|ke|ki|ko|se|aur|karna|karte|kare|lagana|achha|accha|batao)\b

Other Indian languages, for the meaning in tamil type of row:

meaning in (tamil|telugu|bengali|bangla|marathi|gujarati|kannada|malayalam|punjabi)

Devanagari script. RE2 lists Devanagari among its Unicode script classes, so this matches any query containing at least one Devanagari character. It’s the number you’ll want before anyone proposes a Hindi site:

\p{Devanagari}

The full workflow, click by click

One caveat from the same help page: filtering by query can change the report totals, because of data truncation and the omission of anonymised queries. So the filtered numbers are a floor. The real Hinglish demand is a bit bigger than what you’ll see. If you’re new to filtered views, the seven Search Console reports most marketers skip covers regex filters and period comparison more broadly.

Sort the List Before You Touch a Page

Grouped by page, the list usually collapses fast. One explainer catches a cluster: on that store, the lip tint explainer was picking up lip tint kya hota hai, lip tint kya hai, lip tint kya hota h, lip tint meaning in hindi and lip tint in hindi. One page. One fix.

Then split the clusters by what the searcher wants.

The definition rows are where the one-line fix works. The selection rows are commercial and belong to collection pages, which is a different job. Leave them for that work. And the junk is real: that store had meaning in hindi rows for unrelated English words, sitting at positions in the 40s to 70s. Nobody needs a plan for those.

Prioritise by position first, impressions second. Rows on page one or just off it (positions 1 to 15) are the ones a snippet change can move. Rows at 11 to 15, like tinted sunscreen meaning in hindi at 10.9, need the snippet and probably a little more relevance; the position 11 to 100 trap explains why that band is so stubborn.

What to Change: One Hinglish Line Near the Top

Here’s the whole recipe. Add one line near the top of the explainer, inside the opening paragraph or straight after it. Write it the way the dominant query is typed. Follow it immediately with the plain answer.

Illustrative example of the opening of a lip tint explainer.

For meaning in hindi queries, give the meaning in Hindi, because that’s literally what was asked. Romanised first, with the Devanagari word in brackets if you can write it correctly:

A few rules keep this clean.

And don’t spin up a separate lip tint kya hota hai post. You’d be creating a second page for the same question, competing with the explainer that already ranks for it. That’s textbook keyword cannibalisation, self-inflicted. If the page has FAQ schema, add the new question to the markup with the exact visible wording.

On Shopify this is a content edit in the blog post body. No theme change, no app, nothing for a developer.

Title and Meta Description: Only Where the Volume Justifies It

The body line is the low-risk change. Titles are where you need to be careful, because Google has written about exactly this. In June 2022 it described an algorithmic change for titles written in a different language or script from the page. Its examples were Hindi: a Hindi title with an English translation appended, where Google might show only the Hindi part, and a Hindi page with a title transliterated into Latin letters, where Google tries to find a Devanagari alternative from the page.

Apply that to an English explainer. A title written entirely in Hinglish is the mismatch Google describes, so it’s likely to be rewritten or ignored. A title that stays English-led with a short Hinglish phrase added is lower risk. The worst case is that Google trims it back to the English part, and you’ve lost nothing.

These thresholds are judgement calls. Adjust them to the size of your store.

On Shopify, both fields sit in the post’s search engine listing. Shopify lets you enter up to 70 characters in the title and suggests keeping it to 60, and recommends about 160 characters for the description. A Hinglish phrase eats into that budget fast, which is another reason to reserve it for the big clusters. If your titles are already being cut, the truncated title tag fix is worth reading first.

Google’s language-matching note is about titles. I haven’t found an equivalent statement for meta descriptions, and since snippets are mostly built from page content anyway, the body line will usually do more than the meta description does. After any title change, search the query yourself once the page has been recrawled and check what Google actually shows.

When Not to Build a Hindi Site, or Translate Pages

When a Hinglish segment like this turns up, someone always suggests a Hindi version of the store. Usually that’s the wrong call. Here’s how I’d weigh it.

Look at the size first. 3,735 impressions across 207 rows in 28 days is modest. The pages are already on page one. Adding one line to a handful of explainers is an afternoon’s work. A Hindi site means translating every product, collection and policy page, keeping prices, stock and offers in sync, supporting customers in Hindi, and launching a new URL set that Google has to crawl and index from zero.

Then look at the script. These queries are typed in Roman letters. A Devanagari page targets a different script. Google has said it shows Indian-language content for English-typed queries where appropriate, but that’s a statement about results in general. It isn’t a promise that a new Hindi page of yours will outrank the English page that already sits at position 6.

And watch the shortcut. Google’s spam policy on scaled content abuse lists translating among the automated transformations it looks at when pages are mass-produced with little value for users. Translating your own store with care isn’t the same thing. A machine-translated copy of the whole catalogue, published untouched, is a lot closer to that line than you want to be.

For a store that sells across several countries or languages, the bigger question is international structure, and that’s a different conversation. My international SEO work covers when separate language versions pay for themselves, and the four-market international SEO write-up shows what multi-language growth costs in practice.

Hreflang and Devanagari Pages Are a Different Decision

A Hinglish line on an English page doesn’t need hreflang. Hreflang tells Google about localised versions of the same content on different URLs. Google’s hreflang documentation only accepts ISO 639-1 language codes, optionally with a region, and Hinglish doesn’t have one. There’s nothing to annotate. You’ve just written a better English page.

You’ll also read in some Hindi SEO guides that setting lang="hi" helps Google identify a Hindi page. Google says otherwise: it doesn’t use hreflang or the HTML lang attribute to detect a page’s language, and works it out algorithmically from the content. Setting the attribute correctly is still good practice for screen readers and browsers. It just isn’t an SEO signal for language.

If your Devanagari regex shows real demand and the business can serve Hindi-speaking customers end to end, a Hindi version can be worth building. Do it properly:

Don’t auto-redirect visitors to the Hindi version based on what you guess their language is, either. Google’s multilingual guidance warns against it, because it can stop users and search engines from seeing every version.

How to Measure Whether It Worked

Before you edit anything, write down the baseline for the Hinglish segment and for each page you’ll touch: impressions, clicks, CTR and average position, from the regex view.

Then look at the live result. Search the main query on a phone, ideally from India, and see whether your Hinglish line is the snippet. Snippets vary by query, so check two or three variants. Keep this segment separate from your brand reporting too; branded versus non-branded traffic explains why mixing them hides exactly this kind of gap.