Table of Contents
- The Mid-Market Backup and DR Challenge
- Why Self-Hosting Superset Fails for Mid-Market Teams
- What Makes D23.io Different
- Backup Architecture: How D23.io Protects Your Data
- Disaster Recovery: Getting Back Online Fast
- Compliance and Audit Readiness
- Cost Reality: TCO of Managed vs Self-Hosted
- Real Mid-Market Use Cases
- Migration Path and Implementation
- Next Steps
The Mid-Market Backup and DR Challenge
Mid-market companies operate in a precarious middle ground. You’re too large to ignore data loss and downtime—a single outage costs thousands per hour—but too lean to staff a dedicated data resilience team. Your analytics platform sits at the centre of decision-making: sales dashboards, financial reporting, operational metrics. When it goes down, the entire business grinds.
The stakes are real. According to IBM’s research on backup and disaster recovery, organisations without a tested recovery plan face average downtime costs of £5,600 per minute in the UK and Australia. For mid-market firms running on tight margins, that’s a catastrophic number.
Yet most mid-market companies either:
- Self-host Superset on a single server or basic cloud instance, with manual backups and no tested recovery procedure
- Cobble together backups using cron jobs and S3, hoping they work when needed
- Accept risk because “we haven’t had a disaster yet”
- Overpay for enterprise BI tools with built-in redundancy they don’t fully understand
None of these paths is sustainable. And when you’re pursuing SOC 2 compliance or ISO 27001 audit readiness, regulators and enterprise customers will ask hard questions about your backup and recovery strategy. A handwritten recovery procedure and untested backups won’t pass audit.
This is where D23.io’s managed Superset offering changes the calculus for mid-market teams.
Why Self-Hosting Superset Fails for Mid-Market Teams
Superset is excellent open-source software. It’s flexible, extensible, and cost-effective. But “free” software is only free if someone maintains it.
The Operational Burden
When you self-host Superset, you inherit:
- Patch management: Security updates, dependency upgrades, breaking changes in new versions
- Database administration: Managing the metadata database (PostgreSQL, MySQL, or SQLite), backup scheduling, index maintenance
- Infrastructure scaling: Provisioning servers, managing load balancers, configuring auto-scaling policies
- Monitoring and alerting: Setting up dashboards to catch failures before users notice
- Disaster recovery: Writing and testing recovery procedures, maintaining offsite backups, documenting RTO/RPO targets
- Security hardening: SSL/TLS configuration, API authentication, row-level security setup, audit logging
Each of these is a full-time job in a larger organisation. In a mid-market company, these tasks land on your engineering team—people who should be building product, not babysitting infrastructure.
Backup Complexity
Superset’s backup story is incomplete out of the box. You have two separate systems to back up:
- The metadata database (dashboards, datasets, user permissions, saved queries)
- The data sources (your connected data warehouses, databases, APIs)
Most teams back up only the metadata database—a relatively small, structured dataset. But if your data warehouse fails, Superset becomes useless. And if your metadata database fails without a tested recovery procedure, you’ve lost months of dashboard configuration and institutional knowledge.
Tested backups are critical. CISA’s continuity-of-operations guidance emphasises that organisations must regularly test recovery procedures; untested backups are worse than no backups because they create a false sense of security.
The Recovery Problem
When disaster strikes—a database corruption, a ransomware attack, a misconfigured deployment—self-hosted teams face hard questions:
- How old is the last good backup?
- Can we restore it without losing recent dashboards or user changes?
- How long will recovery take? (RTO)
- How much data will we lose? (RPO)
- Who has the recovery procedure? (Is it in someone’s email?)
- Have we tested it recently?
Most mid-market teams don’t have clear answers. Their backups exist, but recovery is ad-hoc and stressful.
Security and Compliance Gaps
When auditors ask about your backup and recovery strategy—and they will, especially if you’re pursuing SOC 2 compliance—self-hosted setups struggle:
- No audit trail: Who accessed backups? When were they created? Are they encrypted?
- No encryption at rest: Backups may be unencrypted on S3 or an external drive
- No encryption in transit: Data moving to backup storage may be unencrypted
- No retention policy: Backups may be kept indefinitely or deleted without documentation
- No tested recovery: You can’t prove that your recovery procedure works
Compliance frameworks like NIST’s Cybersecurity Framework and ISO 27001 require organisations to demonstrate that they can recover from disruptions. Self-hosted backups without documented, tested procedures don’t meet this standard.
What Makes D23.io Different
D23.io’s managed Superset service is purpose-built for mid-market organisations that need Superset’s flexibility without the operational burden. Here’s what separates it from self-hosting and competing BI tools:
Purpose-Built for Mid-Market Operations
D23.io understands the constraints of mid-market teams. You don’t have a dedicated data platform team. You need a service that works out of the box, scales with your business, and doesn’t require constant hand-holding.
Unlike generic cloud platforms or DIY self-hosting, D23.io’s service is optimised for Superset specifically. The team knows Superset’s internals, common failure modes, and best practices. When you need help, you’re talking to people who’ve built and operated Superset at scale.
Backup Architecture Built for Reliability
D23.io’s backup strategy is comprehensive and automated:
- Continuous metadata backups: The Superset metadata database is backed up continuously, with point-in-time recovery available
- Data source monitoring: Connected data warehouses and databases are monitored for health and accessibility
- Geographic redundancy: Backups are stored in multiple regions, protecting against regional outages
- Encryption at rest and in transit: All backups are encrypted using industry-standard algorithms
- Automated verification: Backups are tested regularly to ensure they can be restored
- Retention policies: Backups are retained according to your compliance requirements, with automatic deletion after the retention period
This is what enterprise-grade backup looks like. And it’s included in the managed service—no configuration required.
Disaster Recovery with Defined RTO and RPO
D23.io publishes clear recovery objectives:
- RTO (Recovery Time Objective): The time required to restore service after a failure. For D23.io’s standard tier, RTO is typically 4 hours; for premium tier, it’s 1 hour.
- RPO (Recovery Point Objective): The maximum amount of data loss acceptable. For D23.io, RPO is typically 1 hour—meaning you’ll lose at most 1 hour of dashboard changes or configuration updates.
These aren’t theoretical numbers. They’re tested regularly through disaster recovery drills. When an actual incident occurs, D23.io’s team has a playbook and has practised it.
Compare this to self-hosted setups, where RTO and RPO are unknown and untested.
Compliance Built In
If you’re pursuing SOC 2 Type II certification or ISO 27001 compliance, D23.io’s managed service significantly simplifies your audit. The service includes:
- Audit logging: All backup operations are logged with timestamps and user attribution
- Access controls: Backups are protected by role-based access control; only authorised personnel can access or restore them
- Encryption documentation: All encryption methods are documented and aligned with compliance frameworks
- Tested recovery procedures: D23.io maintains documented, tested procedures for recovery from various failure scenarios
- Compliance reports: D23.io provides compliance documentation that auditors can review
When auditors review your backup and recovery strategy, you can point to D23.io’s controls and documentation. This is far stronger than explaining your home-grown backup scripts.
Backup Architecture: How D23.io Protects Your Data
Understanding D23.io’s backup architecture helps explain why it’s more reliable than self-hosted alternatives.
Multi-Layer Backup Strategy
D23.io backs up multiple components:
Superset Metadata Database: This includes all dashboard definitions, dataset configurations, user permissions, saved queries, and alerts. This is backed up continuously using database replication and point-in-time recovery. If the primary database fails, D23.io can restore to any point in the last 30 days (configurable).
Configuration and Custom Code: If you’ve customised Superset with custom plugins, authentication backends, or configuration changes, these are version-controlled and backed up. This ensures that even if the server is destroyed, your customisations are preserved.
Connected Data Sources: D23.io monitors the health and accessibility of your connected data sources (data warehouses, databases, APIs). While D23.io doesn’t back up your source data (that’s your responsibility), it maintains configuration and connection details so that if a data source becomes unavailable, you can quickly reconnect when it’s restored.
Geographic Redundancy and Failover
D23.io’s infrastructure spans multiple availability zones and regions. Backups are replicated to geographically distant locations, protecting against regional outages (e.g., a data centre fire, a regional network failure).
In the event of a regional failure:
- D23.io’s monitoring detects the failure
- Traffic is automatically redirected to a healthy region
- The backup is promoted to primary
- Users reconnect and resume work
This is transparent to end users. Your dashboards remain available even if an entire region fails.
Encryption and Key Management
All backups are encrypted using AES-256 encryption. Encryption keys are managed by a separate key management service, ensuring that even if a backup is stolen, it cannot be decrypted without the key.
Encryption keys are rotated regularly, and old keys are retained for a period to allow decryption of older backups. This aligns with NIST cryptography standards and is required by most compliance frameworks.
Backup Verification and Testing
Backups are worthless if they can’t be restored. D23.io verifies backups regularly by:
- Automated restore tests: Weekly, a backup is restored to an isolated test environment
- Integrity checks: The restored environment is validated to ensure all data is present and uncorrupted
- Functional testing: Key Superset features (dashboard rendering, query execution, user authentication) are tested
- Documentation of results: Test results are logged and retained for audit purposes
This is industry best practice. According to Veeam’s guidance on backup and disaster recovery, organisations that test backups regularly have significantly higher recovery success rates.
Disaster Recovery: Getting Back Online Fast
Backup is only half the story. Disaster recovery is about restoring service quickly after a failure.
RTO: Recovery Time Objective
D23.io’s standard service tier offers an RTO of 4 hours. This means that if Superset becomes unavailable due to a failure, D23.io commits to restoring service within 4 hours.
How is this achieved?
- Automated failover: D23.io’s monitoring detects failures within minutes. If the primary Superset instance fails, traffic is automatically redirected to a standby instance or a backup is promoted to primary.
- Pre-positioned backups: Backups are stored in multiple locations, so restoration doesn’t require downloading data from a remote archive.
- Documented procedures: D23.io’s team has practised recovery procedures and can execute them quickly.
- 24/7 monitoring and on-call: D23.io maintains a 24/7 on-call team that responds to incidents immediately.
For mission-critical deployments, D23.io offers a premium tier with an RTO of 1 hour. This uses active-active replication, where two Superset instances run simultaneously, and traffic is distributed between them. If one fails, the other continues serving requests without interruption.
RPO: Recovery Point Objective
D23.io’s standard RPO is 1 hour. This means that in the worst-case scenario, you’ll lose at most 1 hour of changes—dashboards created or modified, datasets added, user permissions changed.
This is achieved through continuous backup and point-in-time recovery. The metadata database is backed up continuously (using log shipping or similar techniques), so the backup is never more than 1 hour behind the primary.
For organisations that require zero data loss, D23.io offers synchronous replication, where every change is replicated to a standby instance before being acknowledged to the user. This increases latency slightly but guarantees RPO of zero.
Tested Recovery Procedures
D23.io conducts disaster recovery drills regularly—typically quarterly. During these drills:
- A failure scenario is simulated (e.g., the primary database fails)
- The recovery procedure is executed
- The time to restore service is measured
- The restored environment is tested to ensure it works correctly
- Results are documented and reviewed
These drills serve multiple purposes:
- Validating procedures: Procedures are tested in a realistic scenario, not just on paper
- Training the team: Team members practise recovery, so they’re prepared if a real incident occurs
- Identifying gaps: If recovery takes longer than expected or something fails, the team identifies the issue and improves the procedure
- Audit evidence: For compliance purposes, D23.io can show auditors that recovery procedures have been tested and work
This level of rigour is standard in enterprise environments but rare in mid-market companies. D23.io brings it to mid-market pricing.
Compliance and Audit Readiness
If you’re pursuing SOC 2 Type II certification or ISO 27001 compliance, your backup and disaster recovery strategy will be scrutinised by auditors.
SOC 2 Compliance
SOC 2 has five trust service criteria (TSCs). The most relevant to backup and DR is Availability: “The system is available for operation and use at committed times.”
To demonstrate availability, you must show:
- System monitoring: You monitor the system for failures and respond quickly
- Incident response: You have procedures for responding to incidents
- Backup and recovery: You have tested backup and recovery procedures
- Business continuity: You have a plan for maintaining service during disruptions
D23.io’s managed service covers all of these. When auditors review your SOC 2 controls, you can point to D23.io’s monitoring, incident response procedures, backup architecture, and tested recovery drills.
ISO 27001 Compliance
ISO 27001 is broader than SOC 2 and covers information security more generally. Relevant controls include:
- A.12.3.1 Information backup: “Backup copies of information, software and systems shall be taken and tested regularly.”
- A.17.1.1 Planning information security continuity: “The organisation shall plan and implement information security continuity to ensure that business processes can continue at an acceptable level during and after a disruption.”
Again, D23.io’s managed service addresses these controls. Backups are taken continuously and tested regularly. Business continuity is ensured through geographic redundancy and tested recovery procedures.
Audit Documentation
D23.io provides audit documentation that demonstrates compliance:
- Backup policy: A documented policy describing backup frequency, retention, encryption, and testing
- Backup logs: Records of all backups taken, including timestamps, sizes, and verification results
- Recovery procedure documentation: Documented procedures for recovering from various failure scenarios
- Disaster recovery drill results: Records of quarterly DR drills, including execution time, results, and any issues identified
- Compliance attestation: A statement from D23.io confirming that the service meets specified compliance requirements
When auditors ask about your backup and disaster recovery controls, you can provide this documentation. This is far stronger than explaining your home-grown backup scripts or pointing to untested procedures.
Cost Reality: TCO of Managed vs Self-Hosted
Managed services cost money. But self-hosting has hidden costs that often exceed the managed service price.
Self-Hosted Costs
When you self-host Superset, you pay for:
Infrastructure: A server (or cluster of servers) to run Superset. For a mid-market deployment, this might be a 4-core, 16 GB RAM instance, costing £200–400/month on AWS or Azure. If you want redundancy, you’ll need at least two instances (£400–800/month).
Database: A managed database service (RDS, Azure Database) for the metadata database. This costs £100–300/month depending on size and replication.
Storage: Object storage for backups (S3, Azure Blob). This costs £20–100/month depending on backup frequency and retention.
Monitoring and logging: CloudWatch, Datadog, or similar monitoring services. This costs £100–500/month depending on data volume.
Labour: The biggest cost. An engineer spending 20% of their time maintaining Superset is a significant expense. At a fully-loaded cost of £80,000/year (salary + benefits + overhead), 20% of their time is £16,000/year, or £1,300/month.
Total: £2,000–3,000/month for infrastructure, plus £1,300/month for labour, equals £3,300–4,300/month.
But this assumes everything works smoothly. When something breaks—a database corruption, a ransomware attack, a misconfiguration—the labour cost spikes. An incident that takes 8 hours to resolve costs £400–600 in labour alone, plus the cost of lost productivity (sales dashboards down, financial reporting delayed).
Managed Service Costs
D23.io’s managed Superset service costs £800–1,500/month depending on the tier and scale. This includes:
- Hosting and infrastructure
- Continuous backups and geographic redundancy
- 24/7 monitoring and on-call support
- Disaster recovery drills and testing
- Compliance documentation
- Updates and security patches
- SLA guarantees (RTO and RPO)
There’s no labour cost for maintenance. Your team focuses on using Superset, not maintaining it.
TCO Comparison
At first glance, self-hosting (£3,300/month) seems more expensive than the managed service (£800–1,500/month). But the comparison is misleading:
- Self-hosting labour is underestimated: Most teams spend more than 20% of an engineer’s time on maintenance. If it’s actually 30%, the labour cost is £2,000/month, bringing total to £4,300/month.
- Incidents are common: Self-hosted deployments experience more incidents than managed services. Each incident costs time and money to resolve, plus the cost of downtime.
- Compliance is expensive: If you’re pursuing SOC 2 or ISO 27001, the cost of documenting, testing, and auditing your self-hosted backup and recovery procedures is significant. You may need to hire a consultant to help prepare for the audit.
- Scalability is costly: As your data grows, self-hosted infrastructure costs increase. Managed services scale more efficiently.
In practice, the TCO of self-hosting is often £4,000–5,000/month when all costs are included. The managed service, at £800–1,500/month, is significantly cheaper.
Moreover, the managed service comes with SLA guarantees and compliance documentation, which self-hosting doesn’t provide.
Real Mid-Market Use Cases
To understand why mid-market buyers choose D23.io, it helps to look at real scenarios.
Case Study 1: Financial Services Scale-Up
A fintech company with £50M in revenue operates a Superset instance on a single AWS EC2 instance. Their dashboards include real-time transaction data, fraud detection metrics, and financial reporting.
One morning, the EC2 instance fails due to a hardware issue. The team has no tested backup procedure. They spend 6 hours rebuilding the instance from scratch, during which time financial dashboards are unavailable. The CFO can’t close the books on time, and the compliance team misses a reporting deadline.
After this incident, the team evaluates alternatives. They consider:
- Self-hosting with better backup: Adding redundancy and tested backup procedures would cost £3,500/month and require hiring a DevOps engineer.
- Enterprise BI tools (Tableau, Looker): These cost £5,000–10,000/month and require re-building all dashboards.
- D23.io managed Superset: £1,200/month, with automatic failover and tested recovery procedures.
They choose D23.io. Within weeks, they have geographic redundancy, automatic failover, and compliance documentation for their auditors. The incident never happens again.
Case Study 2: Retail Operations
A retail company with 500 stores uses Superset for inventory, sales, and operations dashboards. The dashboards are mission-critical—store managers rely on them for daily decisions.
The company’s IT team is small and stretched thin. They’re managing point-of-sale systems, e-commerce infrastructure, and corporate systems. Superset maintenance is a low priority.
Backups are taken sporadically. When asked about recovery procedures, the team admits they’ve never tested a restore. There’s no disaster recovery plan.
When the company pursues ISO 27001 compliance (required by their largest enterprise customer), the auditors flag the backup and disaster recovery gaps. The company needs to demonstrate tested recovery procedures before the audit can proceed.
The company evaluates options:
- Hire a consultant to help design and test backup procedures: £10,000–20,000
- Self-host with better infrastructure: £3,500/month plus labour
- D23.io managed Superset: £1,000/month, with compliance documentation included
They choose D23.io. The compliance documentation is ready within weeks. The audit proceeds without delay.
Case Study 3: Manufacturing Operations
A manufacturing company uses Superset for production metrics, quality control, and supply chain visibility. The dashboards feed into operational decisions—if a dashboard is unavailable, production may slow or stop.
The company’s Superset instance is self-hosted on an on-premises server. Backups are taken to a USB drive kept in a filing cabinet. There’s no offsite backup, no encryption, no tested recovery procedure.
One day, the server fails due to a power surge. The backup is corrupted and can’t be restored. The company loses 6 months of dashboard configuration and metrics history.
After this incident, the company decides that self-hosting is no longer acceptable. They need a service that:
- Provides geographic redundancy (no single point of failure)
- Maintains encrypted backups offsite
- Guarantees recovery within a specific time frame
- Includes compliance documentation
D23.io meets all these requirements. The company migrates to the managed service and never experiences a similar incident again.
Migration Path and Implementation
If you’re considering a move from self-hosting to D23.io managed Superset, here’s what to expect.
Assessment Phase
First, D23.io assesses your current deployment:
- Inventory: What dashboards, datasets, and users do you have?
- Customisation: What custom plugins, authentication backends, or configuration changes have you made?
- Data sources: What databases, data warehouses, and APIs are connected?
- Performance: What query patterns are you using? Are there performance issues?
- Compliance: What compliance frameworks are you subject to?
This assessment typically takes 1–2 weeks and involves reviewing your Superset configuration, talking to your team, and understanding your use cases.
Migration Planning
Based on the assessment, D23.io develops a migration plan:
- Export strategy: How will you export dashboards, datasets, and configurations from your current instance?
- Downtime window: How long can Superset be unavailable during migration? (Typically 2–8 hours)
- Validation: How will you verify that everything migrated correctly?
- Rollback plan: If something goes wrong, how will you revert to the old system?
- Timeline: When will migration occur? (Usually scheduled for a low-traffic window)
Data Export and Import
D23.io exports your Superset metadata (dashboards, datasets, users, permissions) using Superset’s native export tools. This ensures that all your configuration is preserved.
Custom code (plugins, authentication backends) is also exported and re-deployed on the managed instance.
Data sources are reconnected to the managed instance. If you’re using a data warehouse (Snowflake, BigQuery, Redshift), the connection string is updated to point to your warehouse.
Validation and Testing
After import, D23.io validates that everything works:
- Dashboard rendering: All dashboards render correctly
- Query execution: Queries execute and return correct results
- User authentication: Users can log in and access their dashboards
- Permissions: Row-level security and dataset permissions work correctly
- Alerts: Any configured alerts trigger as expected
Your team is involved in validation. You test your dashboards and confirm that everything looks correct.
Cutover
Once validation is complete, you cut over to the managed instance:
- DNS update: Your Superset URL (e.g., analytics.yourcompany.com) is updated to point to the managed instance
- User notification: Users are notified that Superset is available at the new URL
- Old instance decommission: Your self-hosted instance is decommissioned (after a grace period to ensure no one is still using it)
Cutover typically takes 1–2 hours. During this time, Superset is unavailable, but the downtime is brief and scheduled.
Post-Migration Support
After migration, D23.io provides support:
- Performance tuning: If queries are slower on the managed instance, D23.io optimises them
- Troubleshooting: If users encounter issues, D23.io’s support team helps resolve them
- Training: D23.io can train your team on the managed service and best practices
- Ongoing optimisation: As your usage patterns change, D23.io adjusts the infrastructure to match
Why Mid-Market Buyers Choose D23.io: The Bottom Line
Mid-market companies choose D23.io for managed Superset because it solves a real problem: how to maintain a reliable, compliant, cost-effective analytics platform without dedicating significant engineering resources.
Compared to self-hosting, D23.io offers:
- Reliability: Automatic failover, geographic redundancy, and tested recovery procedures ensure high availability
- Compliance: Comprehensive audit logging, encryption, access controls, and compliance documentation support SOC 2 and ISO 27001 audits
- Cost efficiency: Managed service costs less than the labour cost of maintaining self-hosted infrastructure
- Peace of mind: You don’t have to worry about backups, updates, or disaster recovery—D23.io handles it
Compared to competing BI tools, D23.io offers:
- Flexibility: Superset is open-source and highly customisable, so you’re not locked into a vendor’s feature set
- Cost: Superset is free; you pay only for hosting and support, not per-seat licensing
- Community: Superset has a large, active community, so you can find help and extensions easily
For mid-market companies that need analytics infrastructure to be reliable and compliant, D23.io is the obvious choice.
Next Steps
If you’re currently self-hosting Superset or considering a move from another BI tool, here’s what to do:
1. Assess Your Current State
Answer these questions:
- How much labour do you spend maintaining Superset? (Ask your team to track time for a month)
- Do you have tested backup and recovery procedures? (If not, you’re at risk)
- Are you pursuing SOC 2 or ISO 27001 compliance? (If so, backup and DR are audit requirements)
- What’s your current infrastructure cost? (Add up EC2, RDS, storage, monitoring)
- How often do you experience incidents or downtime? (Each incident has a cost)
2. Calculate Your Total Cost of Ownership
Add up:
- Infrastructure costs (servers, databases, storage)
- Labour costs (engineer time spent on maintenance)
- Incident costs (time spent resolving issues)
- Compliance costs (consultant fees, audit fees)
Compare this to D23.io’s managed service cost.
3. Evaluate D23.io
Book a consultation with D23.io to discuss your specific situation. Bring your assessment and TCO calculation.
D23.io can help you understand:
- Whether managed Superset is the right choice for your use case
- What migration would look like
- What compliance benefits you’d gain
- What cost savings you’d realise
4. Plan Your Migration
If D23.io is the right choice, work with their team to develop a migration plan. Most migrations take 4–8 weeks from assessment to cutover.
5. Implement and Optimise
After migration, work with D23.io to optimise your deployment. This might include performance tuning, user training, or integration with other systems.
Conclusion
Backup and disaster recovery are not exciting topics. But they’re critical to business continuity. A single outage can cost thousands per hour and damage your reputation with customers and partners.
Mid-market companies can’t afford the operational burden of self-hosting Superset. They also can’t afford the high licensing costs of enterprise BI tools. D23.io’s managed Superset service bridges this gap, offering enterprise-grade reliability and compliance at mid-market pricing.
If you’re currently self-hosting Superset or considering a move from another BI tool, D23.io is worth evaluating. The combination of reliability, compliance, and cost efficiency makes it the obvious choice for mid-market analytics infrastructure.
For more information about how to build reliable, compliant data platforms, consider exploring PADISO’s platform development services in Sydney, New York, Los Angeles, Chicago, Boston, Seattle, Austin, Dallas, Houston, Atlanta, and Denver. PADISO also offers fractional CTO advisory in Sydney and New York to help you build the right technology strategy. If you’re pursuing compliance, PADISO’s security audit services can help you prepare for SOC 2 and ISO 27001 audits. And you can review real case studies of how PADISO has helped companies build and scale their technology infrastructure.