How it started
The first work was operational. Jelmer was running marketing campaigns for Soar Music Group and needed structured lead data: artist contacts, curator profiles, music industry targets. Eprecisio managed development of the scripts that generated those lists.
Over time, a bigger opportunity became clear. Jelmer's internal team was doing everything manually: placing songs on Spotify playlists across multiple profiles, scheduling releases, managing curator relationships, tracking placements. There was no platform. Every operation ran on spreadsheets, scripts, and direct communication.
The idea was to replace all of it with a proper product. One place to manage playlist placements, schedule song drops across multiple Spotify profiles, and handle the curator workflow. That became the first version of Songplace.
The before state
When Eprecisio started building, there was no existing codebase. No prototype. No partial product. The engagement started from zero.
| Area | Manual process before Songplace | Problem this created |
|---|---|---|
| Playlist placement | Individual messages to curators, tracked in spreadsheets | No visibility into status, placements fell through, no audit trail |
| Song scheduling | Manual coordination across Spotify profiles | Human error on release timing, no multi-profile support |
| Curator management | Email and direct messaging per curator | No centralised view, no workflow, no scale |
| Pre-release validation | Manual checks before each drop | Error-prone, time-consuming, limited to team capacity |
| Campaign tracking | No unified system | No data on what worked, no repeatable process |
Jelmer found Eprecisio through online search and selected the team because the pricing fit the budget. What started as a cost-efficient decision became a 6-year relationship because the delivery followed through.
What we built
The engagement has produced three live platforms, built sequentially as each proved its value and the vision expanded.
Songplace (artist and curator dashboards). The core product is a two-sided marketplace for music promotion. Artists submit songs and request streams or playlist placements through a campaign system. Admins review and accept those campaigns, then publish them to the curator network. Curators review the available campaigns, accept the ones that fit their playlists, schedule songs at specific positions, and earn for delivering streams. The platform manages the full lifecycle from submission to placement to payment.
The scheduling infrastructure is the technical core of Songplace. Songs can be pinned to fixed positions in a playlist. The scheduler monitors those placements on a defined interval and corrects any drift, so a song set to position 3 stays at position 3. Pre-release songs have their own validation layer: the scheduler checks placement validity ahead of release, ensures the song goes live at the right time, and handles the transition from pre-release to released state without manual intervention. Playlists also run scheduled refreshes for reach optimisation.
Curator dashboard. The curator-facing side was originally the entire product. It let Jelmer's internal team at Soar Music Group manage all their playlist operations from one place instead of running them manually across Spotify's native tooling. After proving its value internally, it was rolled out to external curators as a standalone product with its own earning model.
FindSocial (separate engagement). Years of scraping infrastructure became a product. FindSocial is a lead platform pre-populated with verified, structured profiles across music and entertainment: artists, curators, labels, agencies, and influencers. Separately documented.
| Component | What we built | Stack |
|---|---|---|
| Artist campaign dashboard | Song submission, campaign creation, stream tracking | React, FastAPI, PostgreSQL |
| Curator dashboard | Campaign review, playlist scheduling, position management, earnings | React, FastAPI, PostgreSQL |
| Scheduling engine | Fixed-position maintenance, pre-release validation, playlist refresh | Python, Celery, PostgreSQL |
| Payment and earnings | Stripe integration for curator payouts and artist campaign billing | Stripe, FastAPI |
| Spotify integration | Multi-profile playlist management, track metadata, listener analytics | Spotify API, Python |
| Infrastructure | AWS-hosted, CI/CD, monitoring, automated deployments | AWS, ArgoCD, GitHub Actions |
| Admin layer | Campaign approval, curator network publishing, platform management | React, FastAPI |
The hardest part: zero errors in scheduled operations
The scheduler is what makes Songplace trustworthy as a product. If a curator pins a song to position 3 in their playlist and it drifts to position 8 overnight, the platform fails its core promise. If a pre-release song goes live at the wrong time, or does not go live at all, the artist's campaign is damaged.
Multiple scheduled operations run simultaneously across every active campaign on the platform: position maintenance checks, pre-release validations, playlist refresh cycles, and placement status updates. Each one has to run cleanly, handle Spotify API rate limits and edge cases, and recover correctly if something fails mid-cycle.
The work to get to zero scheduled operation errors was not a single fix. It was a long process of covering edge cases one by one: what happens when Spotify returns a rate limit error mid-schedule, what happens when a playlist owner removes a song between check cycles, what happens when a pre-release date changes after a campaign has already been confirmed. Every edge case that reached production became a test case that prevented it from happening again.
The second major challenge was multi-profile Spotify management. Artists and labels often operate across multiple Spotify profiles. Coordinating placements, tracking state, and avoiding conflicts across those profiles required careful data modelling that the initial architecture had to be extended to support properly.
The late 2025 revamp
By late 2025, the platforms had years of accumulated code. The product worked. Feature delivery was slowing. The codebase carried technical debt from years of building fast under real business pressure.
Eprecisio ran a full AI-assisted revamp: systematic code audit, removal of dead and duplicate code, test-first policies applied to every refactored module, and a new delivery process built around structured PRDs and sprint discipline.
One sprint is 2 weeks. One quarter is 6 sprints. Before the revamp, the team would plan a quarter's worth of features and deliver most of them by quarter end. After the revamp, the team completed a full quarter's roadmap in 2 sprints, with better code quality and fewer production issues.
The delivery model did not change because the engineers got faster. It changed because the codebase became predictable. When engineers understand every module, have test coverage on every critical path, and work from clear specs, the compounding of small improvements becomes measurable.
Results
| Metric | Before | After |
|---|---|---|
| Playlist operation reliability | Manual, error-prone | Zero errors on scheduled operations |
| Feature delivery | Quarter roadmap completed in 6 sprints | Quarter roadmap completed in 2 sprints |
| Code stability | Accumulated debt, fragile releases | Test-first policy, structured PRD process |
| Platform scope | Manual operations at Soar Music Group | 3 live products serving external customers |
| Engagement size | 1 engineer managing early scripts | 7 to 8 Eprecisio engineers, full product + infra |
| Release cadence | Variable, constrained by instability | Weekly releases, consistent cadence |
"Ehtisham shows great dedication for any project he works on. He understands the issues at hand even if it's very industry specific."
Jelmer Rotteveel, Founder at Songplace via LinkedIn
"Eprecisio is leading all product development and is highly involved on all levels including product design and strategy. Keeping the platforms stable is the most important factor."
Jelmer Rotteveel, Founder at Songplace via Clutch
What the team can do now
Jelmer's internal operations at Soar Music Group ran entirely on manual processes before Songplace. Today, every campaign, every placement, every curator payout, and every playlist schedule runs through the platform. The internal team that used to spend hours coordinating placements by hand now manages the same volume through a dashboard.
The curator network that was previously a set of individual relationships managed over email is now a structured marketplace with its own earning model. External curators can join, accept campaigns, schedule songs, and get paid, without Jelmer's team being in the middle of each transaction.
Commitment, budget discipline, and delivery speed were what Jelmer called out specifically as differentiators in his Clutch review. That combination, staying inside cost constraints while shipping fast enough to keep up with a product that had real traction, is what kept the relationship growing for eight years.
For how we build and maintain full-stack AI-native SaaS products at this scale, see our Development service.
If you are building a SaaS product and need a team that can own the full technical side from architecture to deployment, book a free 30-minute call.