Skip to content
FAQ Library

Every Question. One Place

100+ answers covering Things at Web and our 16 service lines. Use the search to jump straight to an answer, or pick a topic from the sidebar.

About Things at Web

01What does Things at Web do?
Things at Web is a Sweden-based digital engineering company building web, mobile, cloud, AI, and IoT products for enterprises and growth-stage businesses.
02Where is Things at Web headquartered?
Sockerbruksgatan 7, 53140 Lidköping, Sweden. Legal name: Thingsatweb Sweden AB. Delivery teams across Europe and Asia. Clients across the Nordics, EU.
03How big is the Things at Web team?
21+ specialists across data, design, full-stack engineering, hardware, QA, and creative organised as one accountable delivery team rather than siloed functions.
04Do you own the code and design files after the project ends?
Yes. Full ownership of the codebase, hosting accounts, and design files transfers to you on handover. No vendor lock-in.
05Does Things at Web offer SEO and AEO?
Yes. Every site we build ships with technical SEO, structured data (FAQPage / Service / Organization), sitemap, robots, and Core Web Vitals tuning engineered in from day one, not bolted on afterwards.
06How can I contact the team?
Email kontakt@thingsatweb.se or call +46 70 777 07 27. We respond within one business day with discovery-call options.
07Can you take over a project from another vendor?
Yes. We start with a 1–2 week technical audit, document the system, and propose a stabilisation roadmap before changing production code. We do not rewrite reflexively.

Website Development Services

01How long does it take to build a professional business website?
A B2B lead-generation website typically takes 6–10 weeks from kickoff to launch. Complex web applications or e-commerce platforms with custom integrations run 12–20 weeks. Main variables are unique page templates, third-party integrations, and content readiness.
02What makes a website SEO-friendly from a development standpoint?
An SEO-friendly website loads fast (LCP under 2.5s), uses semantic HTML with a single H1 per page, has clean URL structures, generates valid structured data, and is fully crawlable. It serves unique title tags and meta descriptions on every page and handles canonicalisation correctly.
03Do you build websites on WordPress?
Yes but we build WordPress headless where performance is a priority (WordPress as CMS, Next.js or Nuxt as the front end). For content-heavy sites needing editorial flexibility without developer dependency, this gives you the best of both. We also build traditional WordPress when scope supports it.
04Will my website pass Google's mobile-first indexing?
Every Things at Web site is built mobile-first by default. We test on real devices across iOS and Android, validate tap targets and font sizes, and run Lighthouse mobile audits before any launch. All sites pass Google's mobile usability report in Search Console before handover.
05Can you migrate my existing website without losing SEO rankings?
Yes. We run a full pre-migration crawl, map every URL to its new destination, implement 301 redirects, and verify the sitemap before DNS cutover. We monitor Google Search Console and analytics for 4 weeks post-launch to catch any unexpected drops and act immediately.
06Do you provide website hosting and maintenance?
We recommend and configure your hosting (AWS, Vercel, Cloudflare) rather than reselling it you own the account and pay the provider directly. We offer monthly maintenance retainers covering dependency updates, security patches, performance monitoring, and content changes.
07What is a headless CMS and does my business need one?
A headless CMS stores and delivers content via API, decoupled from how it's displayed. Your marketing team edits in a familiar interface while developers use a modern front-end framework. You need it if you publish to multiple channels or if performance is critical and a traditional CMS is slowing your site.

Enterprise AI & Machine Learning Services

01What is the difference between AI consulting and AI development services?
AI consulting defines the strategy which problems to solve, which data you need, what ROI is realistic. AI development services build and deploy the actual systems. Things at Web does both: we start every engagement with strategy to build the right thing, then move directly into development.
02How much data do we need to start an AI project?
It depends on the problem. Classification and regression models often perform well with 10,000–50,000 labelled examples. Time-series forecasting needs 2–3 years of historical data. For LLM applications using RAG, you can start with a few hundred well-structured documents. We assess your data position in phase one.
03How long does it take to deploy an AI model?
A well-scoped predictive model can go from kickoff to production in 8–16 weeks. NLP or computer vision with custom labelling typically takes 16–24 weeks. LLM-based applications with RAG can often be deployed in 4–8 weeks because they leverage pretrained foundation models. Timeline depends on data readiness.
04Can you integrate AI into our existing software systems?
Yes. We build AI as API-first services that integrate with your CRM, ERP, data warehouse, or custom applications via REST or gRPC. Your existing systems don't need to be rebuilt they consume the AI output through a well-defined interface.
05How do you prevent AI models from giving wrong or biased outputs?
Four layers: data auditing to catch biased training sets, statistical fairness metrics evaluated across demographic subgroups, explainability tools (SHAP, LIME) so every prediction is interpretable, and human-in-the-loop review for high-stakes decisions. No model goes to production without documented bias assessment.
06What is RAG and why does it matter for enterprise AI?
Retrieval-Augmented Generation (RAG) is an architecture where an LLM answers questions by first searching your private knowledge base and using retrieved documents as context. AI answers from your actual policies and data not from training data alone. Eliminates hallucination risk for domain-specific queries.
07Do you offer AI model maintenance after deployment?
Yes. Production deployments include a monitoring dashboard for accuracy, latency, and data drift. We offer managed MLOps retainers covering automated retraining triggers, alert escalation, and quarterly performance reviews. Models that aren't maintained degrade we build infrastructure to ensure they improve.

Cloud Services & Cloud Infrastructure

01What is cloud migration and how long does it take?
Cloud migration is moving applications, data, and infrastructure from on-premises or legacy environments to cloud platforms like AWS, Azure, or GCP. A simple lift-and-shift takes 4–8 weeks. A full enterprise migration covering 20–50 applications typically runs 6–18 months in phased waves.
02How much does cloud infrastructure cost vs on-premises?
Most organisations see 20–35% total infrastructure cost reduction after cloud migration, primarily from eliminating data centre overhead, hardware refresh cycles, and paying for peak capacity year-round. FinOps governance right-sizing, reserved instances, and waste elimination is what makes the savings sustained.
03What's the difference between public, private, and hybrid cloud?
Public cloud means shared infrastructure managed by the provider. Private cloud means dedicated infrastructure, on-premises or hosted, managed by you. Hybrid combines both keeping sensitive workloads on-premises while running scalable workloads in public cloud. Most enterprises end up hybrid by design or necessity.
04How do you secure data in the cloud?
Cloud security is a shared responsibility: the provider secures infrastructure; we secure what runs on it. IAM least-privilege, network micro-segmentation, encryption at rest and in transit, secrets management, and continuous security posture monitoring. Every architecture maps to your specific compliance framework.
05Can you reduce our existing cloud bill?
Yes. We start with a 2-week spend audit, identifying idle resources, oversized instances, uncovered reserved instance opportunities, and orphaned storage. Clients typically see 25–40% bill reduction within 60 days. We then implement tagging governance and budget alerting to prevent costs creeping back up.
06Does our business need multi-cloud?
Multi-cloud means running workloads across two or more providers typically to avoid vendor lock-in, meet data residency, or use best-in-class services. Most mid-market businesses don't need it intentionally operational complexity often outweighs benefits. We advise it only with a specific business or regulatory driver.
07How do you handle disaster recovery?
We design DR architectures based on Recovery Time Objective (RTO) and Recovery Point Objective (RPO). Cloud enables active-passive or active-active multi-region at a fraction of traditional DR costs. Critically, we automate DR testing quarterly an untested DR plan isn't a plan, it's a hope.

DevOps & SRE Services

01What is DevOps and how is it different from traditional IT operations?
DevOps is a cultural and technical practice unifying software development and IT operations around shared ownership of the full delivery lifecycle. Traditional IT operations treats infrastructure as separate slow, manual, change-averse. DevOps automates provisioning, testing, and deployment so code flows from commit to production in minutes.
02What is a CI/CD pipeline and why is it essential?
A CI/CD pipeline is an automated sequence taking code from commit through build, test, security scan, and deployment without manual steps. Manual deployment is slow, error-prone, and inconsistent. Teams with mature CI/CD deploy on demand, with lower failure rates and faster recovery than those relying on manual processes.
03What are DORA metrics and why do they matter?
DORA metrics are four research-backed measures: Deployment Frequency, Lead Time for Changes, Change Failure Rate, and Mean Time to Recover. They correlate directly with business outcomes high DORA performers are 2× more likely to meet commercial goals than low performers.
04What is Site Reliability Engineering (SRE)?
SRE applies software engineering principles to operations problems. Core concepts: defining Service Level Objectives (SLOs), using error budgets to balance feature velocity against reliability, and running blameless postmortems to fix systems rather than assign fault. Reliability becomes proactive and measurable.
05How long does a DevOps transformation take?
A functioning CI/CD pipeline and basic infrastructure as code takes 8–12 weeks for most engineering teams. Full DevOps maturity including observability, SRE, platform engineering, and DevSecOps is a 6–18 month journey depending on team size, technical debt, and culture. Culture is the biggest determinant.
06Can DevOps work in regulated industries like finance and healthcare?
Yes regulated industries arguably benefit most. Automated pipelines produce audit trails manual processes cannot. Infrastructure as code makes every change reviewable and reversible. Automated compliance checks are faster and more consistent than manual gatekeeping. We've implemented FedRAMP, SOC 2, and HIPAA-compliant pipelines.
07What's the difference between DevOps and Platform Engineering?
DevOps is the philosophy and practices. Platform engineering builds internal tools an Internal Developer Platform that make DevOps self-service for product teams. A DevOps team helps individual teams with pipelines; a platform engineering team builds golden paths any team can adopt without DevOps on every ticket.

IoT Development Services

01What is industrial IoT (IIoT) and how does it differ from consumer IoT?
Industrial IoT refers to deployments in industrial settings manufacturing, logistics, energy where reliability, security, and deterministic performance are critical. Consumer IoT tolerates occasional failure; industrial cannot. IIoT uses ruggedised hardware, industrial protocols like OPC-UA and Modbus, and operates with extreme temperatures, vibration, and interference.
02How do you connect legacy machines without built-in connectivity?
Most legacy equipment can be connected via protocol converters or edge gateways reading existing interfaces (serial, Modbus RTU, 4-20mA analogue, PLC outputs) and translating to MQTT or OPC-UA. Where no interface exists, we install vibration, temperature, or power sensors externally. Most machines built after 1990 can be connected without mechanical modification.
03What is edge computing and why does IoT need it?
Edge computing means processing data locally on a gateway or edge server rather than sending everything to cloud. IoT needs it for latency (machine shutdown commands can't wait), bandwidth (plants with 500 sensors produce enormous volumes), and resilience (local processing continues when connectivity drops).
04How do you secure an IoT deployment?
Four layers: device identity (each authenticated with a certificate, not a shared password), communication encryption (TLS 1.3), network segmentation (IoT devices on isolated VLANs with controlled cloud egress), and firmware security (signed OTA updates with rollback). We conduct threat modelling specific to your deployment environment.
05How long does an IoT deployment take?
A pilot one production line or site, one use case takes 8–12 weeks from kickoff to live data. A full enterprise deployment across multiple sites with predictive maintenance and ERP integration runs 6–18 months. The biggest accelerator is clear use-case definition and pre-existing OT/IT network infrastructure.
06What is a digital twin and when do you need one?
A digital twin is a virtual model of a physical asset mirroring real-world state in real time using sensor data. Businesses need them to simulate operating scenarios, perform remote diagnostics without site visits, or train operators on equipment behaviour. Most valuable for expensive, complex assets where downtime is costly.
07Can IoT systems integrate with our ERP or CMMS?
Yes. We build API integrations between the IoT platform and your ERP (SAP, Oracle, Dynamics) or CMMS (IBM Maximo, SAP PM, Infor) to automate work order creation from alerts, update asset records with condition data, and feed production actuals into planning. Typically REST API-based, adding 3–6 weeks to deployment.

Digital Transformation Services

01What does digital transformation actually mean?
Digital transformation is the change of an organisation's operating model using digital capabilities modernised technology, redesigned processes, new data foundations, and behavioural change. It is not a tool purchase. Without process redesign and change management, technology investment delivers a fraction of expected value.
02How long does a digital transformation take?
A focused, single-domain transformation (e.g., finance close, sales operations) takes 6–12 months. A multi-domain enterprise transformation is a 2–4 year programme, executed in 90-day waves with measurable outcomes per wave. Transformations longer than 4 years rarely complete.
03Where should we start?
Start where pain is highest and data is cleanest. The two together produce a fast, visible win that funds and de-risks the next wave. Avoid starting with the most strategic process the change-management lift is too large to deliver early credibility.
04How do you measure success?
Per-wave outcome metrics tied to the business case cost-to-serve, cycle time, throughput, NPS, or revenue per FTE measured before, mid, and after the wave. Tool-adoption metrics are leading indicators, not success.
05Do you help with change management?
Yes. We embed change-management practices into every wave: stakeholder mapping, communication plans, capability building, and incentive alignment. Without behavioural change, the new process and tools are abandoned within 6 months.
06What if our existing systems are too messy to transform?
Most transformations start from messy systems that is normal. We run a 2–4 week landscape audit, classify systems by replace / retire / extend / leave-alone, and build the roadmap around realistic dependencies rather than ideal-state architecture.

Data Analytics & Business Intelligence

01What is a modern data stack?
A modern data stack pairs cloud data warehouses (BigQuery, Snowflake, Redshift) with ELT ingestion (Fivetran, Airbyte), SQL transformation (dbt), and BI on top (Power BI, Tableau, Looker, Metabase). It replaces brittle ETL pipelines with reproducible, version-controlled data flows.
02Do we need a data warehouse, lake, or lakehouse?
Most enterprises start with a warehouse for structured analytics. A lake (or lakehouse like Databricks) is added when you also need to store and process unstructured data text, images, logs and run ML on it. Start with the warehouse, add lakehouse capabilities only when use cases justify the complexity.
03How long does it take to ship a usable dashboard?
A first usable dashboard on top of a clean source can ship in 2–4 weeks. If the source needs cleaning, modelling, or integration, add another 4–8 weeks. The bottleneck is rarely the BI tool it is the data preparation and the agreement on definitions.
04Why don't our existing dashboards get used?
Three usual reasons: the metrics don't match how decisions actually get made, the data behind them is mistrusted, or the dashboard is too slow to load. We diagnose all three with a usage audit before rebuilding.
05How do you handle data quality?
Tests at every transformation step (dbt tests, expectations) plus contracts at the source layer. Anomalies are caught at the pipeline, not by the executive viewing the dashboard. Data quality is engineering, not an afterthought.
06Can you connect our SaaS tools and on-prem systems?
Yes most modern SaaS tools have first-class connectors (Salesforce, HubSpot, Stripe, Shopify, Workday). On-prem systems are connected via CDC tooling, direct database replication, or API extracts. We design the ingestion strategy per source.

Custom Software Development

01When should we build custom software vs buy off-the-shelf?
Buy if the off-the-shelf tool solves 80%+ of your need without breaking your workflow. Build when the process is your competitive advantage, when data shape is unusual, or when the workflow does not fit any vendor's box. Hybrid (configure SaaS + build the differentiator) is often the right answer.
02How long does a custom software build take?
A focused internal tool ships in 8–14 weeks. A customer-facing SaaS MVP runs 12–20 weeks. Larger platforms are delivered in 90-day increments with usable software at the end of each. We do not run open-ended retainers.
03What technologies do you build on?
Front-end: React, Next.js, TypeScript. Back-end: Node.js, Python, .NET, Go. Databases: PostgreSQL, MongoDB, Redis. Cloud: AWS, Azure, GCP. We pick the stack for the problem, not because we resell a particular platform.
04Do you own the IP and code we pay for?
You do. Full ownership of the codebase, hosting accounts, and design files transfers on handover. We do not run vendor-lock-in playbooks.
05What does delivery look like day-to-day?
Two-week iterations with demo and review at the end of each. A single point of contact on our side. Stand-ups optional. Slack/Teams channel shared. You see working software, not status decks.
06Can you take over a build from another vendor?
Yes that is a common entry point. We run a 1–2 week technical audit, document the system, and propose a stabilisation roadmap before touching production code. We do not rewrite reflexively.

Technology Consulting

01What does a typical consulting engagement look like?
2–6 weeks. We talk to your team, look at the systems, and produce a documented set of recommendations with a 90-day action plan. Most engagements include 2–4 weeks of light implementation support to start the first wave.
02Do you do technology audits?
Yes. Stack audits, architecture reviews, vendor-takeover audits, and pre-acquisition technical due diligence are common engagements. Output is a written assessment with severity-ranked findings and recommended actions.
03Will you stay to help implement?
Usually yes we prefer engagements that connect strategy to execution. If we recommend a path, we are happy to ship the first wave so you see it working before scaling internally.
04How is consulting priced?
Fixed scope, fixed price for defined audits and roadmap engagements. Time-and-materials for ongoing advisory.
05Can you do board-level advisory?
Yes independent technology advisory for boards and CEOs. Quarterly retainer model with monthly sessions and ad-hoc input on major decisions.

Application Services

01What does application modernisation actually involve?
Re-platforming legacy systems onto modern stacks (containers, cloud-native services), replacing high-maintenance custom code with managed services, and rebuilding interfaces. Sometimes a strangler-fig replacement, sometimes a lift-and-shift first, then refactor. Strategy depends on risk and ROI.
02Can you take over an existing application from another vendor?
Yes. We run a 1–2 week handover audit document the system, capture tribal knowledge, and propose a stabilisation roadmap before changing production code. The first 90 days are typically reliability and observability work, not new features.
03How do you handle application integration?
API-first by default. We build clear contracts between systems and avoid point-to-point integrations that turn into spaghetti. Where the source system has no clean API, we wrap it. Event-driven patterns where state changes need to fan out.
04What do managed application services include?
Monitoring, incident response, patching, capacity planning, dependency updates, and small enhancements. SLA-backed. Typically delivered with a monthly retainer plus a small change budget for new work.
05Will you replace our existing team?
No we augment. The best outcomes are when your team owns the long-term direction and we provide capacity, specialist skills, and process discipline.

IT Infrastructure Services

01What does an infrastructure engagement cover?
Architecture (network, compute, storage), security baseline, monitoring and alerting, backup and disaster recovery, identity and access, and the operational runbooks that keep it running. We hand over fully documented systems.
02Cloud, on-prem, or hybrid?
Workload-by-workload decision. Predictable steady-state workloads with strict data residency often live well on-prem or in private cloud. Variable or growth workloads benefit from public cloud. Most enterprises end up hybrid by design.
03Do you do networking and security?
Yes segmentation, zero-trust patterns, VPN/SD-WAN, firewall policy, IAM, secrets management, and continuous posture monitoring. Threat modelling is part of every architecture engagement.
04Can you take over infrastructure from another provider?
Yes. We start with a baseline audit and ownership transfer plan accounts, credentials, runbooks, documentation. The first weeks are about removing surprises before changing anything.
05What about ongoing operations?
24/7 monitoring, incident response with defined SLAs, patching, capacity planning, and quarterly reviews. Monthly retainer pricing tied to scope.

Product & Hardware Engineering

01What kinds of products do you engineer?
Connected products (IoT devices, sensors, gateways), industrial automation hardware, point-of-sale and retail terminals, and edge-compute appliances. We have shipped products in production retail and industrial environments.
02Do you do electronics design and PCB layout?
Yes schematic capture, PCB layout, EMC awareness, design-for-manufacture. We work with manufacturing partners for assembly and bring-up. Prototyping done in-house.
03Can you do firmware?
Yes embedded C/C++, Rust, MicroPython, and Linux user-space on edge devices. Secure boot, signed OTA, and observability for fielded devices.
04How do you handle mechanical and enclosure design?
CAD-led mechanical design with prototyping (3D print, CNC) and partner-led tooling for production. We design for serviceability and thermal/EMI realities, not just looks.
05Will you support our product after launch?
Yes sustained engineering, bug fixes, firmware updates, and revision planning. We prefer to stay with a product through the first 2–3 years of fielded life so issues feed back into the design.

QA & Testing Services

01What's the difference between QA and test automation?
QA is the discipline of ensuring quality across the SDLC requirements, design, development, release. Test automation is one tool of QA. Strong QA also includes static analysis, code review, observability in production, and clear acceptance criteria not just automated tests.
02Do you do performance and load testing?
Yes k6, JMeter, Locust for synthetic load. We baseline performance at every release and gate on regression. Real-user monitoring (RUM) in production catches what synthetic cannot.
03Can you test for security?
Yes SAST (static analysis), DAST (dynamic scanning), dependency scanning, and threat modelling. Penetration testing for high-risk systems. We embed security testing in the pipeline so it is a continuous gate, not a quarterly audit.
04How do you handle accessibility?
WCAG 2.1 AA as the default standard. Automated checks (axe-core) in CI plus manual screen-reader testing on critical flows. We treat accessibility as a non-negotiable, not an optional retrofit.
05Should we have a separate QA team or embed QA into engineering?
Embed. Teams that own quality ship higher-quality software faster than teams that throw work over a wall to QA. We help organisations move from gated QA to embedded quality engineering.

UI/UX Design Services

01What does your design process look like?
Discover (interviews, analytics, usability review), define (problem framing, success metrics), design (wireframes, prototypes), validate (usability tests, A/B), and deliver (Figma + design system + working alongside engineering). Iterative, not waterfall.
02Do you build design systems?
Yes token-based design systems (Figma + code) with components, patterns, accessibility baked in. We build systems sized to the org over-investment is as costly as under-investment.
03Can you do user research?
Yes interviews, contextual enquiries, usability tests (moderated and unmoderated), and quantitative analytics. We pick the method for the question, not the question for the method we like.
04How do you measure design success?
Per-engagement outcome metrics task completion, time on task, conversion rate, CSAT, retention, NPS measured before and after. Design that does not move a metric is not done.
05Do you redesign existing products?
Yes typically starting from a usability audit and analytics review, not from a blank canvas. The hardest redesigns are the ones where the existing design is mostly right and needs surgical work, not a rebrand.

Blockchain Development Services

01When does a problem actually need blockchain?
When you need a shared ledger between parties who do not trust each other, when token-based programmability is core to the business model, or when immutability is required for compliance or provenance. Most projects do not need blockchain they need a normal database.
02Which chains do you build on?
Ethereum and EVM-compatible L2s (Arbitrum, Optimism, Base) for most contracts. Solana for high-throughput use cases. Permissioned chains (Hyperledger Fabric) for enterprise consortium use cases.
03Do you audit smart contracts?
Yes and we engage an independent third-party auditor on top of our internal review for anything going to mainnet with non-trivial value. We do not skip audits.
04Can you do tokenisation and on-chain payments?
Yes ERC-20, ERC-721, ERC-1155, and the relevant L2 equivalents. KYC/AML integration where regulated. Stablecoin payments and on-chain settlement.
05What about supply chain provenance use cases?
Provenance is one of the cleaner blockchain use cases but the value depends on the off-chain data being reliable. We design end-to-end including the off-chain data path so the on-chain record is meaningful.

Mobile App Development

01Native iOS, native Android, or React Native?
Native (Swift/Kotlin) when deep platform integration, top performance, or platform-only APIs are central. React Native (or Flutter) when the app is mostly UI with shared business logic and time-to-market matters. We choose deliberately per product, not by ideology.
02How long does it take to build a mobile app?
A focused MVP ships in 12–16 weeks. A full consumer-grade app with offline, push, analytics, and integration runs 16–24 weeks. Subsequent feature waves are 4–8 weeks each.
03Do you handle App Store and Play Store submission?
Yes store listings, screenshots, privacy declarations, review responses, and post-launch update cadence. We also handle App Store Optimisation (ASO) basics.
04What about backends?
We can build it. Most mobile apps need a backend API, auth, push, analytics. We build mobile and backend in the same engagement so the API contract matches what the app actually needs.
05Do you maintain apps after launch?
Yes mobile platforms change quarterly, and an unmaintained app rots fast. We offer monthly retainers covering OS upgrades, SDK updates, store policy compliance, and incremental feature work.