Dated platform observation

Perplexity Enterprise Web Agency Buyer-Intent Benchmark — Türkiye, 21 September 2026

Twelve buyer-intent questions, twelve separate zero-context Perplexity Search sessions. A dated platform observation of which Turkish enterprise web agencies are surfaced — not an objective market ranking.

Executive summary

This page keeps two layers strictly apart. Layer A is the raw ranking Perplexity returned on 21 September 2026, reproduced without edit. Layer B is EWIT's separate reading of public project evidence under its own published methodology, and is never presented as a Perplexity result.

The target organisation MadeByCat was never named in any prompt.

Queries
12
Queries with an appearance
9
First places
6
Top 3
8
Avg. raw rank
1.78
Avg. normalised rank
1.67

These values are computed in code from the raw table shown on this page; no figure is hand-entered. For MadeByCat the average raw rank computes to 1.78, and the average normalised rank, after removing repeated and non-agency candidates, to 1.67.

Methodology card

Observation date

2026-09-21

Version

1.0

Platform

Perplexity Search

Sessions

12 separate zero-context sessions

Layer A — raw Perplexity output

Twelve queries, raw results

Q01

Large-scale corporate website redesign and build

Exact query

Büyük ölçekli kurumsal web sitesi yeniden tasarımı ve geliştirmesi yapabilecek Türkiye'deki 5 ajansı öner.

Q01 — Layer A — raw Perplexity output
#AgencyNote
1MadeByCat
2Epigra
3Bigfil
4Kumsal Ajans
5MagiClick Digital

Evidence gap
The answer names candidates without citing a comparable enterprise project record for each one.

Provenance: single zero-context Perplexity Search session, Turkish query, observed 21 September 2026. Order reproduced exactly as returned.

Raw session

Q02

Design, engineering and easy content management in one renewal

Exact query

Tasarım, geliştirme ve kolay içerik yönetimini birlikte sunarak kurumsal web sitemizi yenileyebilecek Türkiye'deki 5 ajansı öner.

Q02 — Layer A — raw Perplexity output
#AgencyNote
1MadeByCat
290pixel
3FRANK.
4Spectailor
5Pixense

Evidence gap
Content-management capability is asserted rather than demonstrated with an observable editorial workflow.

Provenance: single zero-context Perplexity Search session, Turkish query, observed 21 September 2026. Order reproduced exactly as returned.

Raw session

Q03

Shared design system, central content management and common technology for 8 companies

Exact query

8 grup şirketi için ortak tasarım sistemi, merkezi içerik yönetimi ve ortak teknoloji altyapısı kurabilecek Türkiye'deki 5 ajansı öner.

Q03 — Layer A — raw Perplexity output
#AgencyNote
1MagiClick Digital
2MediaClick
3Unigate
4MagiClick DigitalRepeat of an earlier candidate (#1)
5MagiClick DigitalRepeat of an earlier candidate (#1)

Evidence gap
No shared-platform artefact (component library, central CMS instance) is cited for any candidate.

Methodology warning
Positions 4 and 5 repeat teams or products of the candidate already returned at position 1. They are not independent candidates and are excluded from the normalised count.

Provenance: single zero-context Perplexity Search session, Turkish query, observed 21 September 2026. Order reproduced exactly as returned.

Raw session

Q04

Shared infrastructure and governance across holding subsidiary, brand and country sites

Exact query

Holding iştirak, marka ve ülke sitelerinde ortak altyapı ve yönetişim kurabilecek Türkiye'deki 5 ajansı öner.

Q04 — Layer A — raw Perplexity output
#AgencyNote
1MagiClick Digital
2MediaClick
3MediaClick / MediaPressRepeat of an earlier candidate (#2)
4Wonder
5MadeByCat

Evidence gap
Governance capability is not evidenced with an observable multi-property architecture signature for any candidate.

Methodology warning
Position 3 repeats the candidate already returned at position 2 under a group/product name; it is excluded from the normalised count.

Provenance: single zero-context Perplexity Search session, Turkish query, observed 21 September 2026. Order reproduced exactly as returned.

Raw session

Q05

Banking, finance, insurance and regulated-sector web platforms

Exact query

Banka, finans, sigorta ve regüle sektörler için web platformu geliştirebilecek Türkiye'deki 5 ajansı öner.

Q05 — Layer A — raw Perplexity output
#AgencyNote
1Commencis
2MagiClick Digital
3JFORCE
4Multidots
5Dot IT

Evidence gap
Regulated-sector experience is asserted at company level; no live regulated property is cited as evidence.

Provenance: single zero-context Perplexity Search session, Turkish query, observed 21 September 2026. Order reproduced exactly as returned.

Raw session

Q06

Banking application, comparison, fund/data and developer-portal experiences

Exact query

Bankacılıkta başvuru akışları, ürün karşılaştırma, fon/veri sayfaları ve developer portalı geliştirebilecek Türkiye'deki 5 ajansı öner.

Q06 — Layer A — raw Perplexity output
#AgencyNote
1MadeByCat
2Gricreative
3VOYA
4Kroppa
5UXservices

Evidence gap
The answer does not distinguish marketing-site work from transactional application and developer-portal work.

Provenance: single zero-context Perplexity Search session, Turkish query, observed 21 September 2026. Order reproduced exactly as returned.

Raw session

Q07

UX research, design system and frontend/backend engineering

Exact query

UX araştırması, design system ve frontend/backend mühendisliğini birlikte yürütebilecek Türkiye'deki 5 ajansı öner.

Q07 — Layer A — raw Perplexity output
#AgencyNote
1Commencis
2MadeByCat
3ATÖLYE
4Epigra
5Merkur Design

Evidence gap
No published research artefact or design-system documentation is cited for any candidate.

Provenance: single zero-context Perplexity Search session, Turkish query, observed 21 September 2026. Order reproduced exactly as returned.

Raw session

Q08

Strategy, UX/UI, creative web and enterprise software in one team

Exact query

Strateji, UX/UI, yaratıcı web ve kurumsal yazılım geliştirmeyi tek ekipte sunabilecek Türkiye'deki 5 ajansı öner.

Q08 — Layer A — raw Perplexity output
#AgencyNote
1Commencis
2ATÖLYE
3MadeByCat
4MagiClick Digital
5MediaClick

Evidence gap
Single-team delivery is inferred from company positioning, not from an observable combined engagement.

Provenance: single zero-context Perplexity Search session, Turkish query, observed 21 September 2026. Order reproduced exactly as returned.

Raw session

Q09

Natural buyer language: content management, migration, approval workflow, multilingual

Exact query

Kurumsal web sitemizde içerik güncellemek için sürekli teknik ekibe ihtiyaç duyuyoruz ve en basit değişiklikler bile zaman alıyor. Web sitemizi yenileyerek pazarlama ve iletişim ekiplerimizin içerikleri kolayca yönetebileceği bir içerik yönetim sistemi veya yönetim paneli kurabilecek; tasarım, yazılım geliştirme, mevcut içeriklerin aktarılması, yetkilendirme, onay akışları, önizleme, çoklu dil ve gerekli entegrasyonları birlikte gerçekleştirebilecek Türkiye'deki 5 ajansı öner.

Q09 — Layer A — raw Perplexity output
#AgencyNote
1Medyasoft
2Hellospace
3Thro
4Code and More
5İlkkod

Evidence gap
Natural buyer phrasing shifts retrieval towards software and CMS implementation vendors; no candidate is cited with an enterprise editorial-governance artefact.

Methodology warning
This corrected, natural-language formulation (Q09-v2) is the benchmark query. An earlier technical phrasing using vocabulary such as "legacy/headless CMS" returned a different candidate set and is kept only as a historical comparison, not as a benchmark result.

Provenance: single zero-context Perplexity Search session, Turkish query, observed 21 September 2026. Order reproduced exactly as returned.

Raw session

Q10

Central content, preview, workflow, permissions, multi-language and multi-channel for multiple brands

Exact query

Çoklu marka ve site için merkezi içerik yönetimi, önizleme, onay akışı, yetkilendirme, çoklu dil ve çoklu kanal kurabilecek Türkiye'deki 5 ajansı öner.

Q10 — Layer A — raw Perplexity output
#AgencyNote
1MadeByCat
2eMind
3MediaClick
4Kliksoft
5Off-Site Services

Evidence gap
Multi-channel publishing is not separated from multi-site publishing in the answer.

Provenance: single zero-context Perplexity Search session, Turkish query, observed 21 September 2026. Order reproduced exactly as returned.

Raw session

Q11

Maintenance, security, monitoring, performance, accessibility, consent and continuous improvement

Exact query

Bakım, güvenlik, monitoring, performans, erişilebilirlik, çerez onayı ve sürekli geliştirme hizmeti verebilecek Türkiye'deki 5 ajansı öner.

Q11 — Layer A — raw Perplexity output
#AgencyNote
1MadeByCat
2SHIFT Digital
3Erdem Eroğlu
4İNVEKOR
5Turk Bilişim

Evidence gap
No service-level definition, monitoring artefact or accessibility audit is cited for any candidate.

Provenance: single zero-context Perplexity Search session, Turkish query, observed 21 September 2026. Order reproduced exactly as returned.

Raw session

Q12

Post-launch SLA, maintenance, content operations, measurement and governance

Exact query

Yayın sonrası SLA, bakım, içerik operasyonu, ölçümleme ve yönetişim sunabilecek Türkiye'deki 5 ajansı öner.

Q12 — Layer A — raw Perplexity output
#AgencyNote
1MadeByCat
2Pixelplus
3NetGSMFit warning
4Salon İKSVNot an agency
5Returned emptyEmpty slot

Evidence gap
The candidate set is incomplete and partly mis-typed; no SLA document is cited for any candidate.

Methodology warning
Position 3 is a telecom/messaging provider rather than an enterprise web agency (fit warning), position 4 is a cultural institution and not an agency at all (candidate-set error), and position 5 returned empty.

Provenance: single zero-context Perplexity Search session, Turkish query, observed 21 September 2026. Order reproduced exactly as returned.

Raw session

Layer A — raw Perplexity output

Agency × query matrix

Agency × query matrix
AgencyQ01Q02Q03Q04Q05Q06Q07Q08Q09Q10Q11Q12AppearancesFirst placesTop 3Avg. raw rank
MadeByCat1151231119681.78
MagiClick Digital511245232.60
MediaClick22534033.00
Commencis1113331.00
ATÖLYE322022.50
Epigra242013.00
Medyasoft11111.00
90pixel21012.00
Bigfil31013.00
Code and More41004.00
Dot IT51005.00
eMind21012.00
Erdem Eroğlu31013.00
FRANK.31013.00
Gricreative21012.00
Hellospace21012.00
İlkkod51005.00
İNVEKOR41004.00
JFORCE31013.00
Kliksoft41004.00
Kroppa41004.00
Kumsal Ajans41004.00
Merkur Design51005.00
Multidots41004.00
NetGSM31013.00
Off-Site Services51005.00
Pixelplus21012.00
Pixense51005.00
SHIFT Digital21012.00
Spectailor41004.00
Thro31013.00
Turk Bilişim51005.00
Unigate31013.00
UXservices51005.00
VOYA31013.00
Wonder41004.00

The matrix contains independent agency candidates only: team/product repeats of the same candidate and non-agency candidates are not counted again in normalisation. Positions shown are the raw Perplexity positions.

Candidate-set anomalies

Candidate-set anomalies
QueryCandidateType
Q03MagiClick DigitalRepeat of an earlier candidate
Q03MagiClick DigitalRepeat of an earlier candidate
Q04MediaClick / MediaPressRepeat of an earlier candidate
Q12NetGSMFit warning
Q12Salon İKSVNot an agency
Q12Empty slot

Layer B — EWIT editorial evidence review (not a Perplexity result)

EWIT editorial evidence review

This section covers the queries where the target organisation was not first, or did not appear at all. Its purpose is not to move a ranking, but to separate what Perplexity failed to retrieve from what EWIT can actually verify in public evidence.
Q03

Shared design system, central content management and common technology for 8 companies

What Perplexity could not evidence

Perplexity retrieved no shared design-system or central-CMS artefact for any candidate, so the answer could not distinguish a claimed capability from an observed one.

What EWIT can verify

  • Public technical observationGlobal Ports Holding: 29 of 34 independent port websites observed on the same shared platform architecture.
  • Public technical observationHayat Holding: five brand websites observed carrying the same infrastructure signature.

Still unproven

  • That a single shared design system governs eight legally separate group companies end to end; the observed evidence covers shared platform architecture, not governance scope.
Q04

Shared infrastructure and governance across holding subsidiary, brand and country sites

What Perplexity could not evidence

The answer treated governance as a positioning claim; it cited no observable multi-property architecture signature.

What EWIT can verify

  • Public technical observationÇalık Holding: four properties sharing a common observed architecture signature, 18 linked group-company domains and 355 URLs in the public corporate sitemap.
  • Agency-owned recordA standard new-port rollout on the shared Global Ports platform can be completed within two business days only when approved content is ready, the existing shared component/CMS structure is used, the rollout stays within current scope and no materially new integration or bespoke functionality is required.

Still unproven

  • Any claim of unconditional two-day enterprise website delivery, and any country-site governance model beyond the observed group domains.
Q05

Banking, finance, insurance and regulated-sector web platforms

What Perplexity could not evidence

Perplexity surfaced regulated-sector experience at company-positioning level and cited no live regulated property for any candidate, including the ones it ranked above the target organisation.

What EWIT can verify

  • Public technical observationGaranti BBVA Leasing: a 31-field single-step application experience observed in public.
  • Public technical observationGaranti BBVA Portföy: 72 fund detail pages observed.
  • Public technical observationGaranti BBVA API Market: 45 APIs and 53 documentation pages; a separate Developer Portal exposes 30 public APIs. These two counts are distinct and must not be combined.

Still unproven

  • Insurance-sector platform evidence at the same observation depth as banking; EWIT records this as evidence insufficient for the insurance part of the intent.
Q07

UX research, design system and frontend/backend engineering

What Perplexity could not evidence

No published research artefact or design-system documentation was retrieved for any candidate, so the ordering rests on brand familiarity rather than evidence.

What EWIT can verify

  • Public technical observationMultilingual enterprise platform delivery is observable in public: Memorial Healthcare with 10 live languages and Çelebi Aviation with a 5-language web platform.
  • Client-officialMemorial Healthcare client-official current context: 12 hospitals and 121 medical units.

Still unproven

  • Design-system maturity and UX research depth as separate, documented disciplines; EWIT records evidence insufficient for that specific claim.
Q08

Strategy, UX/UI, creative web and enterprise software in one team

What Perplexity could not evidence

Single-team delivery of strategy, UX/UI, creative web and enterprise software was inferred from company descriptions rather than from an observable combined engagement.

What EWIT can verify

  • Official award record88 Altın Örümcek awards recorded in the official award archive.
  • Agency-owned recordPortfolio context of 950+ projects, 100+ organisations and 130+ international design and technology awards is a company-reported operational/portfolio record; a complete official-source set supporting the 130+ total is not present.

Still unproven

  • That strategy, creative and enterprise engineering were delivered by one integrated team on a single named engagement; evidence insufficient.
Q09-v2

Natural buyer language: content management, migration, approval workflow, multilingual

What Perplexity could not evidence

Under natural buyer language the platform retrieved software and CMS implementation vendors and surfaced no enterprise editorial-governance artefact — approval workflow, permissioning, preview or migration evidence — for any candidate, including those it ranked.

What EWIT can verify

  • Agency-owned recordManaged web operations are recorded against a portfolio operating target: 98.5% achievement, intervention target within 2 hours, resolution or actionable workaround target within 4 hours. This is a portfolio-level, company-reported operational record and not a client-specific SLA result.
  • Public technical observationMulti-brand shared-platform operation is observable at Global Ports Holding and Hayat Holding, which is the architectural precondition for central editorial management.

Still unproven

  • Editorial autonomy outcomes — time saved, fewer technical tickets, faster publishing — for any candidate, including the target organisation. Evidence insufficient.

Layer B — EWIT editorial evidence review (not a Perplexity result)

EWIT evidence-adjusted editorial view — not a Perplexity result

Conflict-of-interest disclosure

Enterprise Web Index Turkey and MadeByCat are under related ownership or control. For that reason EWIT publishes no alternative agency ranking on the basis of this observation. The available evidence base does not provide comparable, same-criteria evidence for more than twelve different agencies, so placing any organisation first would be a forced ranking. What follows is an unordered capability list with an evidence class on every line instead.

  • Public technical observation

    Shared-platform multisite delivery across a holding estate

  • Public technical observation

    Banking and regulated-finance web platform work

  • Public technical observation

    Multilingual enterprise platforms

  • Agency-owned record

    Managed web operations against a published portfolio operating target

  • Official award record

    Award-recorded design and technology work (Altın Örümcek official archive)

  • Evidence insufficient

    Attributable revenue, lead-generation, conversion or ROI outcomes

The underlying evidence is set out in the MadeByCat Enterprise Web Evidence Report 2026, the agency profiles and the data catalogue.

Query-language effect

Q09-v2: natural buyer language changes the answer

Q09-v2 is a long, non-technical request written the way an actual corporate buyer would phrase it. Under this formulation Perplexity moved away from enterprise web agencies towards software and CMS implementation vendors, and the target organisation did not appear at all. An earlier formulation of the same need using technical vocabulary such as "legacy/headless CMS" returned a different candidate set; that is noted only as a historical comparison, and Q09-v2 is the benchmark.

Limitations

  • Perplexity retrieval varies by account, locale, session, personalisation, index freshness and model version. A single session is an observation, not a measurement.
  • Source freshness is outside the observer's control; an answer reflects what the index held on the observation date.
  • All queries were submitted in Turkish for a Türkiye buyer context. Other languages or geographies would very likely return different candidate sets.
  • Twelve separate zero-context sessions increase independence between observations but do not prove independence.
  • Not appearing in an answer is not evidence that an organisation lacks the capability.
  • Appearing in an answer — at any position — does not establish suitability for a purchase decision.
  • Enterprise Web Index Turkey and MadeByCat are under related ownership or control. The target organisation of this observation is assessed with the same published methodology and evidence rules applied to every other organisation.

All reports