BharatMatrimony.com
Pan-India flagship
- Sessions at start
- 0
- Ranking keywords
- 0
- Tracked set
- 59 keywords
/ CLIENT CASE STUDY
ORGANIC SEARCH
Ranking India's largest matchmaking network, with one lever only.
Four properties. No on-page access. No backend access. An off-page-only mandate, and 2,406 SERP positions gained in five and a half months.
SCROLL
0Keywords tracked
0Improved position
0→0Avg. SERP position
0Ranking keywords
/ 01 · THE CLIENT
India's largest matchmaking group, four properties deep.
Where a Western dating product optimises for discovery, matrimony optimises for a different search behaviour entirely: families searching with precise, declared filters, community, caste, language, region, profession, marital history. Those filters become queries. Those queries become landing pages.
TeluguMatrimony.com
Andhra & Telangana
- Sessions at start
- 0
- Ranking keywords
- 0
- Tracked set
- 28 keywords
TamilMatrimony.com
Tamil Nadu · caste-dense
- Sessions at start
- 0
- Ranking keywords
- 0
- Tracked set
- 46 keywords
KeralaMatrimony.com
Kerala · life-stage heavy
- Sessions at start
- 0
- Ranking keywords
- 0
- Tracked set
- 75 keywords
Why this portfolio is a hard SEO problem
- Four brands, one entity, overlapping intent. Authority built for one must not cannibalise another, link targeting has to be surgical about which property owns which community term.
- A vast, thin long tail. 54,017 keywords already ranking at the start. Most sat on pages two to five: indexed, existing, and invisible.
- Entrenched, well-funded competition. Mature domains with two decades of accumulated link equity. No easy authority gap to walk into.
- Genuinely seasonal, culturally specific demand. Volume clusters around wedding seasons; the vocabulary is regional. Generic English keyword strategy does not survive contact with this market.
/ 02 · THE BRIEF
Three constraints. None of them moved.
The mandate was simple to state and difficult to execute: raise rankings on a defined set of commercial matchmaking keywords across four properties. What made it unusual was the operating envelope.
01
No on-page access
We could not edit page titles, meta descriptions, headings, body copy, internal linking, or any element of the templates serving the community landing pages.
Consequence The single highest-leverage fix in most SEO programmes, aligning a page's on-page signals with the query it targets, was unavailable.
02
No backend or CMS access
Crawl configuration, site speed, structured data, canonicalisation, redirect handling, XML sitemaps and index management all sat behind an access boundary we did not cross.
Consequence Technical issues identified during audit could be reported but not remediated. Any technical drag persisted for the full engagement.
03
An inherited toxic backlink profile
A high volume of low-quality backlinks pointed at the properties from other matrimony sites, a well-known pattern in this category, where reciprocal and scraped linking is endemic.
Consequence Disavow control sat outside the engagement. New authority had to be built loudly enough to be heard over it.
Strip out on-page, technical and disavow, and one lever remains. The question stops being “what should we fix?” and becomes “how much work can a single link be made to do?”
The framing that set the strategy
/ 03 · THE STRATEGY
Interest-based matchmaking, used as a link strategy.
Generic link building would not have worked here. The properties already had enormous link volume, much of it toxic. So we borrowed the client's own product logic, match on shared interests, community and background, and made it the editorial spine of every off-page asset.
/ 01
Editorial angle
Every asset built around a real cultural topic a publisher would genuinely run, Rajasthani bridal jewellery, Sindhi wedding rituals, Iyengar partner-search customs.
/ 02
Anchor discipline
The brand message and the target query carried in the anchor text, in context, inside content that earned the link on its own merit.
/ 03
Destination precision
Links pointed at the specific community landing page serving that query, never the homepage. This is where most link programmes in this category leak value.
/ 04
Format spread
Long-form guest posts, short-form articles, designed infographics, community Q&A and structured citations, so the profile grew in a shape that looked earned, not placed.
Why the long tail was the right target
Head terms: “matrimony”, “matrimonial site”, “marriage bureau”, are saturated, expensive, and already contested by the client's own brands. The winnable demand sits one layer down, in queries that declare community, region and life-stage.
Critically, the client had already built landing pages for these queries. The content existed. The templates existed. What did not exist was any reason for Google to trust those pages over a competitor's. That is an authority problem, and authority was the one thing we were permitted to build.
Median monthly search volume of the tracked set: 170. Deliberately so, built to win many small qualified queries, not a handful of vanity head terms.
| Family | Example queries | Why |
|---|---|---|
| Community & caste | vishwakarma matrimony kongu vellalar matrimony nair matrimony brides | Highest intent, lowest ambiguity, direct match to an existing landing page |
| Life-stage & marital history | divorced ezhava brides second marriage matrimony widow matrimony | Under-served by competitors, emotionally decisive, strong conversion |
| Profession & region | defence matrimony grooms nair doctor grooms marriage bureau jammu | Narrow volume but near-zero competition and fast movement |
The link architecture
Built in tiers, so authority flowed through a structure rather than pooling on a single layer of low-value placements.
Editorial tierGuest posts on real publications, written to a cultural angle
Carries the exact-match anchor and the strongest trust signal
Tier 2Web 2.0 properties built on contextual content
Reinforces the editorial layer and diversifies referring domains
Tier 3Profile links pointing at the Tier 2 assets
Pushes crawl and equity into Tier 2 so it is discovered and counted
Citation layerBusiness listings, classifieds, bookmarks, image submissions
Entity consistency and a natural-looking profile shape
Competitor-gapPlacements on domains already linking to direct competitors
Closes measurable authority gaps on contested community terms
/ 04 · EXECUTION
6,232 off-page assets, in 5.6 months.
The same scope ran on each of the four properties, which turned the engagement into a controlled comparison across four different query universes.
0Off-page assets builtacross four properties
0Assets per propertyidentical scope on each
0Surfaces workedwebsite · app stores · YouTube
Website programme
Subtotal 1,125 per property
YouTube search
Subtotal 220 per property
App-store optimisation
Subtotal 213 per property
Working method
Keyword-first, not link-first. Each cycle began with a ranking analysis: which community terms sat on pages two and three, which landing page served each, which competitor outranked it and why. Link targets came from that gap analysis, not from a placement inventory.
One anchor, one destination, one query. Every placement mapped to a single target keyword and a single URL before commissioning. This is what made keyword-level attribution possible at all.
Content written to the community, not the keyword. Guest posts briefed as genuine cultural explainers. The anchor sat inside content that stood on its own, which is why placements survived on real publications rather than link farms.
Monthly review against a fixed tracked set. A defined keyword set per property, tracked throughout, so movement was measured on a stable basis rather than a shifting sample.
/ 05 · PORTFOLIO RESULTS
Rankings moved decisively. Traffic moved modestly.
That gap is the story of this engagement, and we take it apart rather than hide it.
0Ranking keywords54,017 → 69,787
0Keywords in top 1012,557 → 15,050
0Keywords in top 2020,131 → 25,027
0Keywords in top 35,866 → 6,315
0Organic sessions1,101,139 → 1,154,780
0Additional sessionsper month, portfolio-wide
Total ranking keywords per property
12 May vs 31 October 2023. Every property grew by at least a fifth. Source: Ahrefs.
BharatMatrimony+35.0%
TeluguMatrimony+21.0%
TamilMatrimony+22.2%
KeralaMatrimony+34.8%
12 May 202331 Oct 2023
Monthly organic sessions, May to October 2023
Every property peaked in July or August and gave ground through September and October, the exact window of the August 2023 core update and the October 2023 core and spam updates. This is a SERP-wide pattern, not a property-specific one.
MAYJUNJULAUGSEPOCT
Bharat 516k → 560k
Telugu 158k → 175k
Tamil 208k → 202k
Kerala 219k → 218k
AUG + OCT 2023 CORE UPDATES
Reading the traffic number honestly
What it is not
A 4.9% traffic gain is not a headline result, and presenting it as one would be dishonest.
What it is
The lagging indicator of an authority programme measured at 5.6 months, through two core updates, with no on-page or technical support, on a base of over a million monthly sessions.
What moved instead
2,406 SERP positions across 208 tracked keywords, page-one presence up from 75 to 164 keywords, and 15,770 net new ranking keywords.
What comes next
Ranking gains in the 4-10 band convert to traffic when on-page signals are aligned to them. The read-out sets out what that unlock looks like.
6.1
Bharat
Bharat
Matrimony
The flagship, and the strongest performer on every ranking measure. Bharat carried the widest query universe, pan-India community terms plus life-stage queries with no regional anchor, which gave the link programme the most surface area to work with.
- Ranking keywords grew fastest in the portfolio, +35.0%, adding 7,785 keywords, more than the other three combined.
- The 4-10 band expanded by almost half: 3,581 → 5,181, +44.7%. This is the band that converts to traffic next.
- Absorbed the deepest recoveries,
pentecostal matrimony98→10,muslim divorce matrimony72→8. - Traffic gained through the volatility, +8.4%, the strongest of the four, peaking at 588,461 sessions in August.
+35%Ranking keywords22,266 → 30,051
+26%Top-10 keywords7,518 → 9,451
+8%Organic sessions516,212 → 559,797
| SERP band | 12 May | 31 Oct | Change |
|---|---|---|---|
| Positions 1-3 | 3,937 | 4,270 | +8.5% |
| Positions 4-10 | 3,581 | 5,181 | +44.7% |
| Positions 11-20 | 3,154 | 4,318 | +36.9% |
| All ranking keywords | 22,266 | 30,051 | +35.0% |
6.2
Telugu
Telugu
Matrimony
The best traffic performer of the four and the most stable through the update window. Telugu's tracked set skewed heavily toward Andhra and Telangana community terms, kapu, kamma, reddy, velama, viswabrahmin, where competition is thin and movement is fast.
- Strongest traffic growth in the portfolio, +10.8%, from 157,881 to 174,854 monthly sessions.
- Ranking keywords up 21.0%, with the 11-20 band growing 21.0%, a large pool of near-miss terms.
- Top-3 count moved sideways: 709 → 697. Gains landed in the 4-10 and 11-20 bands rather than at the very top of page one.
- Deepest single recovery,
sindhi matrimony brides49→18, off the back of a Sindhi wedding-rituals guest post.
+21%Ranking keywords12,466 → 15,088
+7%Top-10 keywords1,876 → 2,010
+11%Organic sessions157,881 → 174,854
| SERP band | 12 May | 31 Oct | Change |
|---|---|---|---|
| Positions 1-3 | 709 | 697 | −1.7% |
| Positions 4-10 | 1,167 | 1,313 | +12.5% |
| Positions 11-20 | 1,597 | 1,932 | +21.0% |
| All ranking keywords | 12,466 | 15,088 | +21.0% |
6.3
Tamil
Tamil
Matrimony
The strongest top-3 performance of the four, and the property where the editorial tier did the most visible work. Tamil's query universe is dense with caste-specific terms, iyer, iyengar, mudaliyar, nadar, kongu vellalar, maruthuvar, each with its own landing page.
- Best top-3 growth in the portfolio, +12.4%, 692 → 778 keywords in positions one to three.
- Ranking keywords up 22.3%: 10,765 → 13,160.
- Two keywords went from page four to the top three,
maruthuvar brides37→2 andmaruthuvar caste matrimony22→1. - Traffic finished slightly down, −2.7%. Peaked at 227,378 in July; the sharpest core-update sensitivity of the four.
+22%Ranking keywords10,765 → 13,160
+19%Top-10 keywords1,859 → 2,203
−3%Organic sessions207,917 → 202,270
| SERP band | 12 May | 31 Oct | Change |
|---|---|---|---|
| Positions 1-3 | 692 | 778 | +12.4% |
| Positions 4-10 | 1,167 | 1,425 | +22.1% |
| Positions 11-20 | 1,597 | 2,008 | +25.7% |
| All ranking keywords | 10,765 | 13,160 | +22.2% |
6.4
Kerala
Kerala
Matrimony
The smallest property by traffic and the second-fastest by keyword growth. Kerala's tracked set was the largest of the four at 75 keywords, concentrated on ezhava, nair, thiyya and vishwakarma community terms plus a strong divorced- and second-marriage cluster.
- Ranking keywords up 34.8%: 8,520 → 11,488, nearly matching the flagship's growth rate off a much smaller base.
- The 11-20 band grew 40.2%: 1,226 → 1,719, the fastest near-miss expansion in the portfolio.
- Life-stage queries moved hardest,
defence matrimony grooms55→7,nair divorced grooms49→6. - Traffic held roughly flat, −0.6%, having peaked at 228,373 in August.
+35%Ranking keywords8,520 → 11,488
+6%Top-10 keywords1,304 → 1,386
−1%Organic sessions219,129 → 217,859
| SERP band | 12 May | 31 Oct | Change |
|---|---|---|---|
| Positions 1-3 | 528 | 570 | +8.0% |
| Positions 4-10 | 776 | 816 | +5.2% |
| Positions 11-20 | 1,226 | 1,719 | +40.2% |
| All ranking keywords | 8,520 | 11,488 | +34.8% |
/ 07 · KEYWORD PERFORMANCE
208 keywords. Every single one improved.
The set was defined at the start and never changed, so it gives a clean read on the programme's effect, uncontaminated by keywords entering or leaving the index.
0Positions gained11.6 per keyword
0Improvedzero keywords lost position
75 → 0On page one+119%
0 → 0In the top threefrom none at baseline
Where the keywords ended up
Distribution of the 208 tracked keywords by SERP band. The 51+ and 21-50 bands emptied into 4-10 and 1-3.
1-3
053
4-10
75111
11-20
4133
21-50
6111
51+
310
Start · 12 MayEnd · 31 Oct
Movement profile
0gained 1-4 positions36%
0gained 5-9 positions29%
0gained 10-19 positions19%
0gained 20-39 positions11%
0gained 40+ positions5%
74 of 208 keywords: 36% of the tracked set, gained ten or more positions. The distribution is what you would expect from authority-led work: a broad base of modest gains on already-competitive terms, and a long tail of very large recoveries on pages that had authority added to them for the first time.
The sixteen largest moves
Grey marks the start position, green the end position. Scale runs 100 → 1.
- pentecostal matrimonyBharat · 720/mo+88
- muslim divorce matrimonyBharat · 110/mo+64
- rajasthani marriageBharat · 320/mo+64
- mudiraj brideBharat · 320/mo+62
- divorced women in indiaBharat · 480/mo+60
- nri marriage bureauBharat · 320/mo+60
- knanaya matrimonyBharat · 140/mo+59
- defence matrimony groomsKerala · 210/mo+48
- ezhava second marriage groomsKerala · 20/mo+43
- nair divorced groomsKerala · 260/mo+43
- thuluva vellalar matrimonyKerala · 110/mo+41
- indian christian matrimonyBharat · 140/mo+39
- divorced ezhava bridesKerala · 140/mo+36
- maruthuvar bridesTamil · 480/mo+35
- maharashtrian brideBharat · 590/mo+34
- kashyap rajput mehra bridesBharat · 210/mo+32
/ 08 · WORKED EXAMPLES
Four placements, traced from commission to ranking change.
Aggregate numbers are only persuasive if the mechanism underneath them is visible. Each example shows the full chain: target query, landing page, publication, the live article carrying the anchor, and the position change that followed.
What the four have in common
Every one started off page one.
Start positions were 10, 47, 76 and 55. These were pages Google knew about and had no reason to rank.
Every one ended on page one.
End positions were 3, 20, 21 and 7, three of them inside the top ten.
Every one was a cultural explainer.
Wedding rituals, bridal jewellery, partner-search customs, second-marriage narratives. The editorial angle is what earned the placement.
Domain authority mattered less than relevance.
The DA 17 placement on theindianpreneur.com produced a 48-place gain, larger than the DA 88 placement on nairaland.com. Topical fit beat raw authority in every comparison we can make here.
/ 09 · LOCAL SEARCH & OTHER SURFACES
106 store locations, audited end to end.
Alongside the four websites, Matrimony.com operates a national network of physical branches. All 106 locations across four regional clusters were brought to a consistent standard.
0Verified104 of 106
0With imagery105 of 106
0Phone approved97 of 106
0Map pin correct103 of 106
Issues found and resolved
- Nine phone numbers rejected by Google. Four in the Karnataka/AP/Telangana/Maharashtra/Gujarat cluster and five in Kerala, the largest single data-quality issue in the network, and a direct barrier to call volume from local search.
- Three listings with an incorrect map pin. Latitude/longitude corrections in Tamil Nadu, Kerala and West Bengal. A wrong pin undermines the proximity signal that drives local pack inclusion.
- Two verifications outstanding at close. One location in Peelamedu still in Google's processing queue, one in the West Bengal/Odisha cluster unverified.
- One listing without imagery. 105 of 106 listings carried photographs by the end of the engagement.
App-store optimisation
App-store rankings respond to off-store signals. The programme built those directly at the Play Store and App Store URLs, alongside a community Q&A layer on Quora aimed at install-intent questions.
| App listing | Placement | DA |
|---|---|---|
| TeluguMatrimony Google Play | justpaste.it | 91 |
| TeluguMatrimony Apple App Store | penzu.com | 76 |
| TamilMatrimony Google Play | tumblr.com | 86 |
| BharatMatrimony Google Play | twitter.com | 94 |
| KeralaMatrimony Google Play | storeboard.com | 62 |
YouTube search optimisation
Sixty videos per property reviewed and optimised each month, title and description strategy, keyword research, tag and thumbnail recommendations, with an off-video link layer built to support them.
- 60 videos optimised per property, per month
- 120 blog commenting links
- 20 Web 2.0 submissions
- 20 video submissions
Caveat Video-level performance data was not part of the reporting scope shared with us, so this documents delivery rather than outcome. Any future engagement should instrument YouTube Analytics before the work starts.
/ 10 · READ-OUT
What worked. What didn't.
Worked
- Anchor and destination precision. Mapping one keyword to one URL before commissioning a placement is the single practice most responsible for these results, and what made keyword-level attribution possible at all.
- Cultural editorial angles. Guest posts framed as genuine community explainers landed on real publications and held. Promotional angles would not have.
- Targeting the near-miss band. Concentrating on keywords already sitting in positions 11-30 produced the fastest movement. Pages already had relevance; they needed authority.
- Relevance over raw domain authority. The DA 17 topically-perfect placement outperformed the DA 88 general placement, consistently enough to be a planning rule, not an anecdote.
- Running the same scope across four properties. Identical delivery on four different query universes turned the engagement into a controlled comparison.
Didn't
- Traffic did not follow rankings within the window. +4.9% portfolio-wide against 2,406 positions gained. Rankings in the 4-10 band produce far less traffic than positions 1-3.
- Top-3 growth was the weakest measure. Portfolio top-3 keywords grew only +7.7%, and Telugu's top-3 count fell slightly. Getting from position four to position two generally requires on-page alignment we could not perform.
- Two core updates absorbed the gains. August and October 2023 both landed inside the window. All four properties peaked in July or August and gave ground afterwards.
- The toxic backlink profile was never addressed. Disavow control sat outside the engagement. Every new link had to outweigh an inherited negative signal that remained in place.
- YouTube and ASO outcomes were not instrumented. Delivery is documented; impact is not, because the measurement access was not in scope.
The programme did the job it was scoped to do. The constraint that defined it, no on-page, no backend, is also the ceiling it hit.
The honest conclusion
What the next phase should be
The engagement built a large population of keywords sitting in positions 4-20. Across the portfolio that band grew from 14,265 to 18,712 keywords. That is stored value: pages one alignment step away from the top of page one.
- 01Unlock on-page on the top 200 near-miss pages
Title, meta, H1 and internal-link alignment on the community landing pages now sitting in positions 4-10. The highest-return work available, currently blocked by access, not by budget.
- 02Run a technical audit with remediation rights
Crawl efficiency, page speed, structured data and canonical handling on template-generated community pages. At this scale, template fixes compound across thousands of URLs.
- 03Clean the inherited backlink profile
A disavow programme against the toxic matrimony-network links, so newly built authority is not spent offsetting a negative baseline.
- 04Instrument the full funnel before the next cycle
Search Console separated per brand, GA4 configured for organic-only reporting, plus YouTube and app-store analytics.
- 05Convert the 11-20 band with content, not just links
The 11-20 band grew to 9,977 keywords portfolio-wide. With content editing rights, that band is the fastest traffic unlock in the portfolio.
/ 11 · METHODOLOGY
How every figure was produced.
| Measure | Source | Basis |
|---|---|---|
| Organic sessions (domain level) | Ahrefs | Snapshot at 12 May and 31 Oct 2023 |
| Monthly organic sessions | Ahrefs | Month-end readings, May-Oct 2023 |
| Ranking keyword counts & SERP bands | Ahrefs | Snapshot at 12 May and 31 Oct 2023 |
| Tracked keyword positions | Screpy (paid rank tracker) | Fixed keyword set, monthly reads |
| Keyword search volumes | Ahrefs | India monthly estimates |
| Store listing status | Google Business Profile | Full audit of 106 locations |
| Placement domain authority | Third-party DA metric | As recorded at time of placement |
Known limits
Endpoint comparison. Unless stated otherwise, “before” is 12 May 2023 and “after” is 31 October 2023.
Fixed tracked set. The 208 keywords were defined at the start and not changed, so no survivorship bias enters the position statistics. This set was selected as a target list, read it as a measure of the programme's effect on its own targets, not as a random sample.
Third-party estimates. Ahrefs traffic figures are modelled estimates, appropriate for direction and relative change, not server-side session counts.
Algorithmic context. The August 2023 core update and the October 2023 core and spam updates all fell inside the window. No attempt has been made to isolate their effect from the programme's.
Attribution. Because no on-page, technical or content work ran in parallel, ranking movement on targeted keywords is attributable to the off-page programme or to SERP change. There is no third workstream competing for credit.
Note on figures Several headline percentages in the original reporting deck do not reconcile with the underlying tool data printed alongside them. Every figure here has been recalculated from source values rather than carried over.