Service · Consultancy

Google Cloud consulting for organisations that want their cloud set up properly.

Architecture advice and delivery for clients who run on Google Cloud or are considering it. We design scalable, secure and cost-efficient GCP environments — and build them ourselves too. Vendor-independent: we know AWS and Azure just as well, and we won't choose Google Cloud if something else fits better.

GCP Landing ZoneWell-Architected reviewsCost optimisationBigQuery & Vertex AI

Advising and building on Google Cloud — no separate tracks.

Google Cloud consulting at Appfront is not an abstract architecture engagement detached from delivery. We have been working with Google Cloud for years for clients who run production workloads on the platform — web applications, customer portals, data platforms on BigQuery, an integration layer via Pub/Sub and, increasingly, AI solutions built around Vertex AI and the Gemini API. The conversations about architecture, security and cost are held with the same people who then deliver the work.

We are a Dutch builder-led agency with a cloud practice. Our position is deliberately independent: no reseller, no licence quotas, no incentive to sell you more Google Cloud services than you need. If Azure would serve you better for a particular workload, we'll say so. The starting point is always your situation, not a vendor's portfolio.

Google Cloud consulting is an umbrella over several more specific engagements. If you want to take a broader approach, you can start with broad digital consulting. If you have just decided to move to the cloud, you often begin with cloud migration support. And for clients still choosing a platform: see also the parallel pages on AWS consulting and Azure consulting — the same principles, different tools.

What we mean by Google Cloud consulting.

The question "do you do Google Cloud consulting?" covers very different areas of work. Below are the six that come up most often in our practice — in different proportions in every engagement.

Architecture

Cloud architecture design

A GCP design that fits your workload: which services, how they are distributed across projects and folders within the Resource Hierarchy, with what network segmentation via Shared VPC and what redundancy approach via multi-zonal or multi-regional deployments. Not a reference architecture from a whitepaper, but a design that fits your traffic pattern, team size and growth assumptions.

Review

Well-Architected reviews

A structured review along the pillars of the Google Cloud Architecture Framework: operational excellence, security, reliability, cost, performance and sustainability. We dig into the Terraform code, IAM bindings, VPC topology and runtime metrics in Cloud Monitoring. At the end you get a prioritised list with concrete next steps.

Costs

Cost optimisation

GCP bills tend to climb in a way that rarely feels linear to the business. We analyse where the money is really going — over-provisioned Compute Engine, expensive cross-region egress, unused static IPs, missing Committed Use Discounts, poorly placed BigQuery slots — and build an approach to make the worthwhile savings stick.

Security

Security & IAM

Folder and project hierarchy with Organization Policies, identity via Cloud Identity, machine identity via Workload Identity Federation (no more long-lived service account keys), encryption via Cloud KMS, network segmentation via VPC Service Controls and Private Service Connect, audit logging via Cloud Audit Logs, and threat detection via Security Command Center. We work from a clear information security policy — not from a checklist applied after the fact.

Migration

Migrating to Google Cloud

From on-premises, from AWS or from a SaaS context. We map the existing landscape, choose the right migration strategy for each workload (rehost, replatform, refactor or replace), and build the migration path without bringing the business to a halt. For the broader context, see cloud migration support.

Modernisation

Containers, serverless & events

If you already run on Google Cloud but have a lot of "lift-and-shift" workloads, modernising can pay off: containers via GKE or serverless via Cloud Run, event-driven flows via Pub/Sub and Eventarc, short transactions on Cloud Functions, an API layer via API Gateway or Apigee. Not as an end in itself, but where it makes things simpler, cheaper or more robust.

Three ways we deliver Google Cloud consulting.

Depending on where your GCP practice stands and which decision you have on the table. We decide together in the first conversation which format fits — an engagement often starts with a review and grows into an implementation phase.

Compact project · fixed sprint budget

Google Cloud Well-Architected review

A structured scan of your existing GCP environment along the pillars of the Google Cloud Architecture Framework. We look at IAM bindings, Organization Policies, VPC topology, cost drivers, observability via Cloud Logging and Cloud Monitoring, backup strategy and deployment pipelines. At the end you get a findings report plus a prioritised list — not an endless document, but concrete improvements your team can start on tomorrow.

Architecture FrameworkIAM auditCost scanQuick wins
Mid-sized project · fixed sprint budget

GCP Landing Zone and architecture design

A design for a new Google Cloud environment or a redesign of an existing one. A GCP Landing Zone following Google's Enterprise Foundations Blueprint: organisation and folder hierarchy, project vending, Shared VPC with hub-and-spoke, identity baseline in Cloud Identity, baseline policies via Organization Policy Service and an IaC approach using Terraform with the Cloud Foundation Toolkit. At the end you have an environment in which your teams can build securely and predictably.

Landing ZoneShared VPCTerraform/CFTOrg Policies
Ongoing engagement · sprint-based

Guided implementation and modernisation

Strategy without execution stays theory. In this engagement we support not only the architecture decisions but also the implementation: migrating workloads, CI/CD via Cloud Build and Artifact Registry, modernising to GKE, Cloud Run or Cloud Functions, and observability via Cloud Logging, Cloud Trace and Cloud Profiler. We stay on board until it works in production and your team can carry on building independently.

CI/CDGKECloud RunObservability

Which Google Cloud services we use in practice.

Google Cloud now has hundreds of services. Honestly, no one knows them all in detail, and neither do we. What we do is deliberately focus on the services we work with regularly in client projects, and build deep practical expertise in those.

For compute, Cloud Run is often the starting point: managed containers, scale to zero, low management overhead, and suitable for the vast majority of modern web applications and APIs. We use GKE where the team has Kubernetes maturity: Autopilot for managed nodes, Standard for finer-grained control. Cloud Functions for event-driven or short-running code. For data, the focus is on BigQuery (discussed separately below), complemented by Cloud SQL, Spanner for globally distributed transactional systems, AlloyDB for PostgreSQL-compatible OLTP, Firestore for document stores and Cloud Storage for object storage.

For security and identity, we use Cloud Identity, Identity Platform or Firebase Authentication for customer-facing authentication, Cloud KMS and Secret Manager, and VPC Service Controls as a perimeter around sensitive data stores. For AI workloads, Vertex AI and the Gemini API are the two services we work with most: Vertex AI as the managed platform for training, fine-tuning, deployment and evaluation, plus the Gemini API for multimodal foundation models. Document AI for data extraction from unstructured documents, Vision AI for image recognition. For event architecture, Pub/Sub, Eventarc and Workflows; for API management, API Gateway and Apigee. For observability, Cloud Logging, Cloud Monitoring, Cloud Trace, Cloud Profiler and Error Reporting. For an AI project on Google Cloud, see AI development.

BigQuery as the heart of a data platform.

A large share of our Google Cloud projects revolve around data. BigQuery is the reason many organisations choose GCP in the first place: a serverless, columnar warehouse that scales without clusters to maintain, with SQL as the interface and a pricing model that separates storage and compute. That is fundamentally different from a classic warehouse that you pay for 24/7.

What makes BigQuery even more interesting is that it is set up to be AI-friendly. BigQuery ML lets you train models with SQL statements directly on your data. BigQuery integrates with Vertex AI for heavier models and with Gemini for multimodal queries, so you can ask questions of text, image and structured data in the same session. BigQuery supports Apache Iceberg as an open table format, and with BigLake you can run the query engine on data in Cloud Storage, Amazon S3 or Azure Blob Storage without copying it. That opens the door to data lakehouse architectures that are not tied to a single cloud.

In practice, we design data platforms in which Pub/Sub or Datastream handles ingestion, Dataflow runs the stream and batch transformations, dbt or Dataform structures the modelling and BigQuery forms the central warehouse. Looker or Looker Studio for BI on top, and where needed Vertex AI Pipelines for heavier ML workflows.

What you concretely get from a Google Cloud project.

No vague deliverables, but artefacts your cloud team, security officer and business owners can put to use straight away.

  • Architecture documentA readable overview of the chosen Google Cloud services, how they work together and which trade-offs were made. Not just a diagram, but the reasoning behind it.
  • Infrastructure as codeA Terraform repository (with Cloud Foundation Toolkit modules where appropriate), reproducible and linked to your CI/CD via Cloud Build or GitHub Actions. No manual console clicking.
  • GCP landing zone baselineFolder and project hierarchy, a project vending approach, Shared VPC with hub-and-spoke, baseline Organisation Policies and a runbook for how new projects and workloads are added securely.
  • Identity and security baselineCloud Identity setup, IAM bindings following least privilege, Workload Identity Federation instead of service account keys, VPC Service Controls, a Cloud KMS strategy and Security Command Center integration.
  • Cost overview and optimisation planInsight through Cloud Billing reports and BigQuery exports of the billing data, plus concrete measures: VM rightsizing, Committed Use Discounts, BigQuery slot reservations and network egress optimisations.
  • CI/CD and observability setupDeployment pipelines via Cloud Build and Artifact Registry or GitHub Actions, plus Cloud Monitoring dashboards, alerting policies, Cloud Trace, Cloud Profiler and Error Reporting.
  • Hands-on build role (optional)If you want, we stay on board as a delivery partner for migration, modernisation or continued development on the new environment.

When Google Cloud consulting makes a difference.

Six situations in which clients come to us for GCP work. If you recognise one, we'd be happy to talk further. A no-obligation conversation is enough to see whether the chemistry is right.

Greenfield

A new environment for a new product

You're starting a new platform and have chosen Google Cloud as the foundation, often because BigQuery, Vertex AI or the developer experience of Cloud Run tipped the balance. You want to get it right first time: a clean folder and project structure, IaC via Terraform, an identity baseline in Cloud Identity, and deployment pipelines. Not a messy proof of concept that later has to be painstakingly tidied up.

Data & AI

A data platform built around BigQuery

You're building a serious data platform and choosing BigQuery as its core, often because serverless pricing, BigQuery ML, Iceberg support and Vertex AI integration fit where your team wants to go. You're looking for a partner who designs ingestion (Pub/Sub, Datastream, Dataflow), modelling (dbt, Dataform), governance (Dataplex) and the BI layer (Looker) as one whole.

Audit

The bill is higher than it should be

You already run on Google Cloud, but the monthly bill has structurally risen higher than the business expected. You suspect overcapacity, expensive cross-region egress, BigQuery queries eating up slot budget, or missing Committed Use Discounts, but the exact picture is missing. A Well-Architected review focused on the cost pillar puts the insight on the table and gives you levers to structurally do something about it.

Compliance

GDPR, DORA, NIS2 or NEN pressure

Your sector sets requirements for data residency, logging and incident procedures: GDPR, DORA for financial service providers, NIS2 for essential and important entities, NEN 7510 for healthcare. The Google Cloud environment must demonstrably comply, and you're looking for a partner who can get both the design and the evidence in order. See also ISO 27001-compliant software development.

AI on GCP

Making Vertex AI and Gemini production-ready

You're experimenting with Vertex AI and the Gemini API and want to take it from prototype to production, with network isolation via Private Service Connect, content filtering, prompt logging that stays within the EU data boundary, cost monitoring per model, and a fallback strategy for model deprecations. We've been through this journey before and know where the sharp edges are.

Scale

The MVP is outgrowing itself

The platform is running in production, customers are enthusiastic, and growth is outpacing what the architecture was designed for. Single-region becomes multi-region, a handful of Cloud Run services become dozens, and the original choices are starting to pinch. You're looking for a partner who can lift it to the next level of maturity.

EU data residency and compliance on Google Cloud.

For European customers, data residency is often not a detail but a prerequisite. Google Cloud offers several EU regions. For Dutch customers, europe-west4 is often a logical choice: the data centre is in the Eemshaven (Groningen), with low latency to the rest of the Netherlands. For disaster recovery, europe-west1 (Saint-Ghislain, Belgium) is a natural partner, with europe-west3 (Frankfurt) as a second good option. Which region fits depends on your customer base, legal context and service availability — not every service runs in every region. On top of that, Google offers an EU Data Boundary for customer and support data, a commitment that data for covered services stays within European regions. It's no silver bullet — data can still leak through misconfigured logging destinations or cross-region features — but the framework for data sovereignty on GCP is more mature than it used to be.

For financial service providers, the DORA regulation has applied since 2025: digital operational resilience, exit strategies for critical ICT providers and third-party risk management. For healthcare providers, NEN 7510 remains the framework. Essential and important entities have obligations under NIS2. And for all personal data, there is the GDPR. We design your Google Cloud setup so that it demonstrably supports these frameworks — with evidence that comes structurally from Cloud Audit Logs, Security Command Center and Access Transparency.

One specific point of attention is vendor lock-in without an exit plan. Some managed services gradually make an organisation hard to move. For every significant service choice, we put the dependency explicitly on the table: what would an exit look like, and do the benefits outweigh the portability you give up? For BigQuery, Iceberg support and BigLake can play a role in an exit strategy: the data sits in an open format that other engines can read as well.

How a Google Cloud consulting engagement works with us.

1

Introduction and scope kick-off

An initial conversation in which we understand where your Google Cloud practice stands, which decision is on the table and what constraints apply — in terms of budget, compliance, team skills and timeline. After that comes a short proposal phase in which we agree which format — review, architecture design or guided implementation — fits best.

2

Research and discovery

We get read access to your Google Cloud organisation and look at Terraform repositories, IAM bindings, Organization Policies, VPC topology, Cloud Billing and monitoring data. We also talk to cloud engineers, your security officer and application owners — not to know everything, but to be able to ask the right questions.

3

Workshops and architecture choices

In a number of focused sessions, we lay out the options: service choices, folder and project structure, network blueprint (Shared VPC versus Private Service Connect), identity and security baseline, data platform approach around BigQuery, and deployment approach. No black-box analysis — your team takes part in the decisions and retains ownership.

4

Plan and business case

We translate the choices into a phased plan covering sequencing, team deployment and investment. It is concrete enough for the board or the finance team, and flexible enough to adjust as your understanding develops.

5

Delivery (optional)

If you wish, we stay on board for the implementation — as a building partner, programme manager or interim cloud engineer. That way, advice and execution are carried by the same people, and the engagement only ends once the Google Cloud environment is running in production and your team can continue building on it independently.

Frequently asked questions.

What clients usually ask us before we start: honest answers, no sales talk.

Are you an official Google Cloud Partner?
No. We are an independent software development agency that works a lot with Google Cloud, but we don't present ourselves as a Google Cloud Partner with a specific specialisation. For some customers, that is actually an advantage: no channel pressure, no partner incentives and no "we have to mention this service because we're a partner". Our value lies in hands-on experience with production workloads on GCP — not in a logo. If you explicitly need a partner relationship, for example for a Google-funded migration programme or a marketplace procurement route, we'll say so honestly and help you find the right party.
Do you also work with AWS or Azure?
Yes. We have hands-on experience with all three major hyperscalers and see no reason to be dogmatic about choosing one. Google Cloud is often a strong choice for organisations that rely heavily on data and AI — BigQuery as the warehouse, Vertex AI and Gemini as the AI platform, Cloud Run as a developer-friendly compute layer. If you're deep in a Microsoft stack, Azure may make more sense; if you need the breadth of the AWS ecosystem, AWS. We also advise on where a hybrid or multi-cloud setup makes sense — and where it only adds unnecessary complexity.
How do you handle vendor lock-in on Google Cloud?
Deliberately. Not by avoiding all managed services — that often costs more in build and maintenance than it saves — but by explicitly naming the dependency with every significant service choice. For some workloads, deep GCP integration through Cloud Run, Pub/Sub and Spanner is perfectly fine; for others, a container approach on GKE is the wiser choice precisely because it could run elsewhere on Kubernetes with limited changes. For data, Iceberg support and BigLake play a role in the exit strategy — your data sits in an open format that other engines can read too. We weigh this up case by case.
What is a GCP Landing Zone, and do we need one?
A GCP Landing Zone is a standardised baseline setup following Google's Cloud Adoption Framework and the Enterprise Foundations Blueprint: organisation, folder and project hierarchy, a project-vending approach, a Shared VPC or hub-and-spoke network blueprint, baseline policies via the Organization Policy Service, an identity layer in Cloud Identity, and an approach to logging, monitoring and billing. For organisations that work with Google Cloud on a structural basis, it is almost always the sensible choice — restructuring a messily built GCP organisation after the fact is far more expensive than putting a good landing zone in place up front. For very small organisations with a single GCP project, a full landing zone is overkill; in that case we recommend a lighter baseline that can grow later.
Which Google Cloud services do you typically work with?
Compute: Cloud Run, GKE and Cloud Functions. Data: BigQuery at the core, plus Cloud SQL, Spanner, AlloyDB, Firestore, Cloud Storage and Bigtable. Edge: Cloud CDN and Cloud Load Balancing. Identity: Cloud Identity, Identity Platform or Firebase Authentication, and Workload Identity Federation for machine identity. AI: Vertex AI, the Gemini API, Document AI and Vision AI — see also our AI development page. Events and APIs: Pub/Sub, Eventarc, Workflows, API Gateway and Apigee. Observability: Cloud Logging, Cloud Monitoring, Cloud Trace, Cloud Profiler and Error Reporting. That's the core; around it we work with countless other services whenever the situation calls for it.
Where do we store data — in which Google Cloud region?
For Dutch customers, usually europe-west4 (Eemshaven, Groningen), with europe-west1 (Belgium) or europe-west3 (Frankfurt) as a secondary region for disaster recovery. For customers with a Scandinavian focus, europe-north1 (Finland) may be a better fit. What choosing a region does not automatically solve is data leakage via logging, AI or management services. The Google EU Data Boundary is a serious commitment that customer and support data stays within the EU for the covered services, but it is no excuse for not checking the configuration. That has to be set up separately through Organization Policies, correct KMS key allocation, VPC Service Controls and deliberate choices around multi-region storage. We do this setup as part of a security baseline.
Do you work with Cloud Build or GitHub Actions?
Either, with a slight preference for GitHub Actions: the developer experience is stronger, the community is larger and the Google Cloud integration through the official google-github-actions is mature. Workload Identity Federation makes it easy to deploy from GitHub Actions to GCP without long-lived service account keys. For organisations that want to keep everything in a single pipeline within GCP, especially when Artifact Registry, Binary Authorization and Cloud Deploy are deeply integrated, Cloud Build is a perfectly good choice. It's not a religion; our preference is for whichever tooling the client's developers are comfortable with.
Why do organisations choose BigQuery over another warehouse?
The combination of serverless pricing (storage and compute billed separately), a mature SQL engine, BigQuery ML for running models directly on the data, BigLake and Iceberg support for open table formats, and tight integration with Vertex AI and Gemini for multi-modal queries. For organisations moving towards an AI-first approach, BigQuery often feels more natural than a classic warehouse. For traditional BI workloads, Snowflake or Databricks may fit just as well or better, depending on your existing tooling and team skills. We're not dogmatic; the choice between warehouses is an open question in every engagement.
How do you differ from a Google Cloud-only consultancy?
A GCP-only firm often knows the platform inside out. We differ in three ways. First, we're vendor-independent, so choosing Google Cloud is an outcome of the analysis, not the starting point. Second, we also build applications, not just infrastructure, which avoids handovers. Third, our pricing suits mid-market budgets. If you've already chosen GCP and only need platform-specific optimisation, a GCP-only firm may suit you well. If you want advice and delivery to stay connected, we occupy a different niche.
What determines the cost of a Google Cloud consulting engagement?
Mainly the scope and the complexity of the existing environment. A Well-Architected review for a small, straightforward Google Cloud project is a different engagement from an assessment of a multi-folder organisation with dozens of workloads, Shared VPCs and complex Cloud Identity federations. We work with a fixed sprint budget per phase, so you know where you stand, and we only extend it if you decide to move on to a next phase.
Do you work with our internal cloud or IT department?
Almost always. In the first phase we actively reach out to your cloud engineers and security officer to understand what's already running and which sensitivities are at play. During delivery we work with mixed teams wherever possible: your people plus ours, with clearly divided ownership. Knowledge transfer isn't an afterthought: when we leave, your team must be able to manage and further develop the Google Cloud environment independently.
How does a programme start in practice?
With an introductory conversation, no obligation. We listen to your situation and advise which format fits: a review, an architecture design or guided implementation. For a review, we first get read-only access so that we can base our proposal on facts.
Fabian van Dijk Business developer · fabian.vandijk@appfront.nl

Talk to us about your Google Cloud question.

An introductory conversation, with no obligation and no sales pressure. We listen to your situation, ask questions and give you direction you can actually use, even if it turns out that working with us is not the right next step, or that another cloud provider is a better fit.

Edit content