Insights
Notes from the group.
How we build and operate companies: company building, AI operations, data and security, commerce, marketplaces and growth. RSS
Operating note ·
Outbound clicks, destination sessions and leads: what each number proves
A practical checklist for interpreting portfolio referrals: define the event, verify the destination, compare consistent periods and separate a click from an arrival or enquiry.
Announcements ·
Introducing Oryvelon: the group behind our companies
Oryvelon is the company builder that starts, owns and operates MerchNivo, CastLyra, EduRelia, ZodiVela, KeşifAtlası, Noveniq, WeAreMedia and Sinem Keser Beauty Academy. This is why it exists, how it works and the rules it holds itself to.
Company Building ·
What is a digital company builder?
A digital company builder repeatedly starts, owns and operates digital businesses on a shared set of people, technology and standards. Here is how the model works, where it comes from, how it differs from agencies, studios and holding companies, and how Oryvelon applies it.
Company Building ·
Why we don't merge our products into one app
Shared infrastructure is efficient. Shared products usually are not. Why Oryvelon keeps MerchNivo, CastLyra, EduRelia, ZodiVela, KeşifAtlası and its other companies as independent products — and what we share instead.
AI Operations ·
AI explains, verified data decides
How Oryvelon uses language models in products where a wrong answer has real consequences — stock levels in MerchNivo, visa rules in KeşifAtlası, student access in EduRelia — and the engineering rules that keep AI from inventing facts.
Data, Privacy & Security ·
Shared infrastructure, separate data
How a group of companies can share technology without building a cross-product user graph. The data boundaries Oryvelon sets for each company, why marketing consent is not product consent, and what every product plans for from day one.
Company Building ·
How we start a new company: our 10-step protocol
Ideas are cheap; companies are expensive. The ten steps every idea must pass before it becomes an Oryvelon company — from a one-sentence problem to a continue, stop or scale decision — and why the order matters.
AI Operations ·
Cost discipline for AI products
AI features can quietly become the largest line on a product's bill. How Oryvelon keeps AI cost visible and controlled across a portfolio: per-product budgets, model routing, caching, batching, output limits and quarterly tool reviews.
Company Building ·
Company builder, venture studio or holding company? A practical comparison
Company builders, venture studios, holding companies, agencies and funds are often confused. A clear comparison of how each one works, how it earns, what it controls and when each model makes sense — and why Oryvelon chose to be a company builder.
Company Building ·
How we validate a digital business idea before building it
Validation is the cheapest stage of a company's life and the one most often skipped. The questions, signals and small experiments Oryvelon uses to decide whether an idea deserves a real product — and the signals we ignore.
Company Building ·
Phase-zero sites: launching a brand before the product exists
Every Oryvelon company gets a real website before its product is finished. What a phase-zero site is, what it must contain, how it is measured, and why it often becomes a company's most valuable long-term asset.
Company Building ·
Continue, stop or scale: how we make portfolio decisions at 30, 60 and 90 days
Every Oryvelon company has written 30/60/90-day KPIs and an explicit decision at each checkpoint: continue, stop or scale. How the checkpoints work, what evidence we look at, and why stopping is a skill a company builder must practise.
Company Building ·
Designing a recurring revenue loop from day one
A company that must re-win every customer every month is exhausting to run. How Oryvelon designs recurring revenue into each company from the first sketch — subscriptions, usage, credits, licences, repeat purchases and partner loops — with examples from across the group.
Company Building ·
Capital allocation inside a company builder
A company builder's most important job is deciding where money, time and attention go. How Oryvelon allocates resources across its companies: evidence over enthusiasm, visible per-company economics, staged commitments and the discipline to move resources away from what is not working.
Company Building ·
Designing every company so it could stand alone
Any Oryvelon company could one day raise capital, take a partner or be sold. That is only realistic if it can be separated cleanly. The technical, legal and operational choices we make from day one so each company can stand on its own.
Company Building ·
Why every company starts with a source-of-truth document
Before code, design or marketing, every Oryvelon company gets a written source of truth: what it is, who it serves, how it earns, what it will not do and which rules apply. Why this one document saves so much time, what goes in it and how we keep it alive.
Company Building ·
Brand architecture for a multi-company group: parent brand vs. product brands
How Oryvelon relates to MerchNivo, CastLyra, EduRelia, ZodiVela, KeşifAtlası, Noveniq, WeAreMedia and Sinem Keser Beauty Academy as brands. Why we chose independent product brands with a quiet parent, how 'An Oryvelon company' is used, and how names and domains are chosen.
Company Building ·
The operating cadence behind a portfolio of digital businesses
Running several companies well depends less on brilliant strategy than on a steady rhythm: weekly signals, monthly reviews, quarterly checks on access, tools and costs, and checkpoint decisions. The operating cadence Oryvelon uses across its companies.
AI Operations ·
What an AI gateway does — and what it should never do
Every AI feature across Oryvelon's companies goes through a shared gateway for model routing, budgets, prompt registry, safety hooks, logging and fallbacks. What the gateway is responsible for, how it is designed, and the hard line it must never cross: merging products' prompts, knowledge or user data.
AI Operations ·
Model routing: matching tasks to the right AI model
Model routing decides which AI model handles each task, so small models do the simple, high-volume work and larger models are reserved for problems that need them. How we write routing rules per feature, measure the cost/quality trade-off and keep routing a configuration change rather than a rewrite.
AI Operations ·
Prompt versioning and evaluation for production AI
Prompts in production are code that changes behaviour for every user at once, so we register them, version them, test them against fixed evaluation sets and keep a one-step rollback. How a prompt registry works, what goes into a test set and how we decide a new version is ready to release.
AI Operations ·
Structured outputs: validating AI responses before users see them
When an AI feature feeds a screen, a workflow or a database, its output needs a contract. How we use JSON schemas, validation, bounded retries and reference checks to reject malformed or hallucinated responses before they reach users, and what the application does when validation fails.
AI Operations ·
Earning user trust in AI products during onboarding
People decide whether to trust an AI product in its first few minutes. How we design onboarding to set honest expectations, show where answers come from, ask only for the data we need and deliver a first piece of real value quickly — with examples from MerchNivo, KeşifAtlası, EduRelia, ZodiVela and CastLyra.
AI Operations ·
Designing human escalation paths in AI products
Every AI product needs a clear point where the machine stops and a person takes over. How we decide what gets escalated, how the handover works without losing context or leaking data, and how KeşifAtlası, CastLyra and EduRelia each design escalation for very different kinds of risk.
AI Operations ·
AI fallbacks and degraded modes: what a product does when the model fails
Models time out, providers go down and budgets run out. How we design AI fallbacks and non-AI degraded modes for every Oryvelon product, so a failed model call becomes a plainer experience rather than a broken one, and users keep the facts they came for.
AI Operations ·
Latency and caching for AI features: keys, batching and streaming
How we keep AI features fast and affordable across Oryvelon's companies: measuring latency where users feel it, caching with keys built from product, prompt version and input fingerprint, batching work that nobody is waiting for, and streaming only where it genuinely helps the reader.
AI Operations ·
RAG without data leakage: per-product and per-tenant retrieval
Retrieval-augmented generation is the easiest place for AI products to leak data, because a model will happily use anything it is handed. How we build RAG across Oryvelon's companies with separate indexes per product and tenant, access filtering before retrieval, and no shared knowledge between companies.
AI Operations ·
Safety boundaries for consumer AI products: pre and post hooks
Consumer AI products need boundaries that hold even when users push on them. How we define content limits for ZodiVela and EduRelia, and enforce them with pre-request and post-response safety hooks, product-specific rules and human review, rather than relying on a single instruction in the prompt.
Data, Privacy & Security ·
Data retention and deletion: schedules, flows, backups and logs
Keeping data costs money, adds risk and rarely helps. How each Oryvelon company sets its own data retention schedule, builds deletion flows that reach every copy, handles backups and logs honestly, and treats AI prompts, caches and indexes as data that must be deleted too.
Data, Privacy & Security ·
Tenant isolation explained: separate databases, schemas and row-level security
Tenant isolation decides whether one customer's data can ever appear in another customer's screen, report or AI answer. We compare separate databases, separate schemas and row-level isolation, and explain why Oryvelon isolates at two levels: between companies and between each company's customers.
Data, Privacy & Security ·
Child data minimisation in edtech: what to collect, what to refuse and who decides
Edtech products that serve children should collect the least data that lets learning work, and treat the school as the party that decides. How we approach child data minimisation at EduRelia: schools as controllers, consent that makes sense to families, and a written list of things we will not collect.
Data, Privacy & Security ·
Marketing consent vs product consent: why we keep them separate in every company
Agreeing to use a product is not agreeing to be marketed to. How Oryvelon separates product consent from marketing consent, keeps consent records inside each company, and refuses to build cross-brand marketing lists — plus the data model, wording and checks that make the separation hold.
Data, Privacy & Security ·
Least-privilege access for small teams: roles, contractors, 2FA and access reviews
Small teams tend to give everyone admin because it is faster. A practical least-privilege access model for startups and small companies: role-based access, time-limited contractor accounts, mandatory 2FA, quarterly access reviews and a same-day offboarding routine — and how we run it across Oryvelon's companies.
Data, Privacy & Security ·
Secrets management basics: API keys, rotation and keeping credentials out of code
API keys and credentials cause some of the most avoidable security incidents in small companies. The secrets management basics we apply across Oryvelon: keys only on the server, separate keys per company and environment, scoped permissions, scheduled rotation, and a firm rule that secrets never go into repositories, tickets or chat.
Data, Privacy & Security ·
Environments and test data: local, preview, staging and production without real personal data
Most small companies test with copies of production data because it is easy. How we set up local, preview, staging and production environments across Oryvelon's companies, build synthetic test data that behaves like the real thing, and keep production personal data out of every environment except production.
Data, Privacy & Security ·
Domain, DNS and email security for a multi-brand group
How we protect the domains, DNS and email of every Oryvelon company: registration and ownership, registrar locks and auto-renew, one DNS standard, and SPF, DKIM and DMARC on every sending domain. A practical baseline any group running several brands can copy.
Commerce & SaaS ·
What an AI e-commerce employee is, and what it is not
An AI e-commerce employee for Shopify stores reads the store's real data, writes a daily briefing, flags operational signals and suggests actions that a person approves. Here is how we define the role at MerchNivo, where its limits sit, and why human approval is part of the design.
Commerce & SaaS ·
Shopify operations signals: what to watch daily and weekly
The operational signals a Shopify store should watch every day and every week: stock cover, refunds and returns, abandoned checkouts, fulfilment delays, payment issues and discount misuse. How to set sensible thresholds, and how MerchNivo turns these signals into a short, ranked briefing.
Commerce & SaaS ·
Running a real store as a testbed for commerce software
Oryvelon operates Noveniq, a direct-to-consumer technology accessories store, as a real business that also teaches us what commerce software like MerchNivo should do. How we get practical learning from a live store without sharing customer data, and what that arrangement can and cannot prove.
Commerce & SaaS ·
Subscription plus usage pricing for AI SaaS
Why AI software products often need a subscription plus a usage component: the subscription pays for the service and predictable value, usage covers variable AI costs. How we design plans, set fair limits, choose usage units customers understand and avoid surprise bills.
Commerce & SaaS ·
Inventory as source of truth: why AI must never guess stock
Inventory is one of the few numbers in e-commerce where a confident wrong answer causes immediate damage: overselling, missed reorders and broken promises. How we treat inventory as a source of truth, how to sync and reconcile stock across systems, and why MerchNivo's AI never estimates it.
Commerce & SaaS ·
Running a DTC brand as a real operating business
A direct-to-consumer brand is not a website with ads pointed at it. It is a product business with margins, stock, fulfilment, returns and repeat purchase to manage. How we run Noveniq, our technology accessories store, as a proper operating company and what that discipline teaches the rest of the group.
Commerce & SaaS ·
B2B2C licensing to schools: how the model really works
In B2B2C school licensing, the school buys and students and families use the product. That split changes procurement, onboarding, data roles, support and renewal. How EduRelia approaches each stage, and what any edtech founder should design for before signing a first school.
Commerce & SaaS ·
Unit economics per product: AI costs and shared cost allocation
When several companies share infrastructure and an AI gateway, it becomes easy to lose track of what each product really costs to serve. How we calculate unit economics per company and per product, where AI costs belong, and how we allocate shared costs without flattering or punishing anyone.
Marketplaces & Consumer Products ·
Verification in talent marketplaces: identity, portfolio and consent
A talent marketplace is only as trustworthy as its verification. How CastLyra approaches verifying creative talent and the brands that hire them: identity, portfolio ownership, image consent, business legitimacy, contact-detail rules and the trust signals that make each check visible and useful.
Marketplaces & Consumer Products ·
Solving the marketplace cold-start problem
Every two-sided marketplace starts empty: buyers will not come without supply, and supply will not come without buyers. A practical approach to the cold-start problem, built on supply first, a narrow niche and geography, and manual matching, with examples from how we think about CastLyra.
Marketplaces & Consumer Products ·
Marketplace monetisation models: commission, subscription, lead fees and featured listings
Commission, subscriptions, lead fees and featured listings each change how a marketplace behaves, not just how it earns. A practical comparison of the main marketplace monetisation models, the incentives they create, and which combination fits a verified talent marketplace like CastLyra.
Marketplaces & Consumer Products ·
Rules engines for eligibility: why AI explains and verified rules decide
Eligibility questions — can this person apply for this visa, this permit, this programme — need answers that are traceable, versioned and repeatable. How we design a rules engine for KeşifAtlası, where verified rules decide the outcome, AI explains it in plain language, and rule changes are handled as releases.
Marketplaces & Consumer Products ·
Positioning consumer AI as reflection and entertainment, honestly
Some consumer AI products are meant for reflection and enjoyment, not for decisions. How we position ZodiVela — astrology, numerology, tarot and traditional readings — with honest framing, no predictions or advice, careful handling of sensitive moments and design that makes the experience a pleasure rather than a pressure.
Marketplaces & Consumer Products ·
From free test to paid report: designing an honest consumer funnel
A free test followed by a paid report is one of the clearest ways to sell a consumer AI product — if the line between free and paid is honest. How KeşifAtlası and ZodiVela decide what is free, what is paid, how trust is earned before the payment step and how refunds and quality issues are handled.
Marketplaces & Consumer Products ·
Why a services company lives inside a company builder
Oryvelon builds and operates its own companies and does not sell agency services. Client work belongs to WeAreMedia, a separate services company inside the group. Why we structure it this way, where the boundaries sit on data, brand, pricing and people, and what a services business teaches a portfolio of products.
Marketplaces & Consumer Products ·
Services, education and digital products: building a founder-led expertise business
A founder with real expertise can earn from the same craft in three ways: delivering the service, teaching it and packaging it as digital products. How Sinem Keser Beauty Academy combines appointments, courses and digital products into one business with recurring loops, clear boundaries and a founder whose time is protected rather than consumed.
Growth, Search & Measurement ·
Search and AI visibility for companies: how to be the clear, consistent answer
Search and AI visibility for companies comes down to three things: a consistent entity that machines can recognise, genuinely helpful pages that answer real questions, and the discipline to avoid programmatic spam. How we approach discoverability in Google and AI assistants across Oryvelon and its independent companies.
Growth, Search & Measurement ·
llms.txt and structured data: the machine-readable layer of a company website
llms.txt and structured data give search engines and AI assistants a clean, unambiguous description of who you are and what you publish. What llms.txt, schema.org Organization and subOrganization, Article, FAQPage markup and sitemaps each do, and exactly what oryvelon.com publishes for a group of independent companies.
Growth, Search & Measurement ·
UTM standards across brands: one naming convention for a group of companies
UTM standards across brands keep campaign data readable when a group runs several companies and founder sites. The naming convention we use for utm_source, utm_medium and utm_campaign, how links between Oryvelon, its companies and founder sites are tagged, and the governance that stops analytics from filling up with near-duplicate labels.
Growth, Search & Measurement ·
Analytics without surveillance: privacy-respecting measurement for a group of companies
Analytics without surveillance means measuring what a business needs to decide — traffic, conversion, retention, cost — without building profiles of people. How we set up privacy-respecting analytics across Oryvelon: a separate property for every company, aggregate metrics first, consent handled per audience, and no cross-brand user tracking.
Growth, Search & Measurement ·
Founder authority sites: how personal sites relate to a group and its companies
Founder authority sites give the people behind a company a home for their own work, voice and credibility. How ugurkeser.art and sinemkeser.com.tr relate to Oryvelon and its companies, why links between them stay contextual rather than sitewide, and how to keep a personal site, a group site and company sites clearly distinct for people and machines.
Growth, Search & Measurement ·
The quarterly tool review: how we keep software subscriptions, costs and access under control
A quarterly tool review is a short, recurring check of every software tool and subscription a business pays for or depends on. How we run it across Oryvelon's companies: a single inventory, a named owner for every tool, duplicate detection, cost attributed to the right company, and an access review that removes what nobody needs.