Case stories

Advertising Intelligence

Guided audience building and campaign planning powered by behavioral signals.

Hoppr partnered with DEVVIA to build Darwin, an AI-assisted planning studio that ingests millions of signals in BigQuery, recommends high-performing segments, and orchestrates multi-phase campaigns.

8.5M+Signals/hr

Behavioral and platform events processed through BigQuery.

-34%Media waste

TURF simulations optimize reach before spend is locked.

15 minCampaign launch

AI assistant builds multi-phase plans end-to-end.

Challenge

Manual segmentation and spreadsheets slowed modern ad ops.

Hoppr’s analysts stitched together spreadsheets, siloed BI dashboards, and manual reach studies to craft campaigns. Audience definitions drifted between systems, and validation required hours of overlap analysis run offline.

  • User-level behavioral data was trapped in bespoke exports with no shared taxonomy.
  • Reach and overlap calculations required offline tooling that delayed approvals.
  • Creative assets, prompts, and telemetry lived in disconnected tooling.
Approach

Governance-first delivery pod execution.

DEVVIA deployed a cross-functional pod-cloud architects, data scientists, and product designers-to ship a governed Google Cloud platform with embedded AI assistants, Cloud Run analytics, and reusable planning modules.

Behavioral signal platform

Unified millions of user-level signals into BigQuery with privacy-safe contracts, standardized taxonomies, and streaming ingestion.

  • Data Contracts enforce naming, consent tags, and expiration.
  • Segment APIs hydrate the campaign builder and external DSPs.

AI-assisted campaign studio

Built an AI assistant that guides users through segment creation, suggests best-fit audiences, and assembles phased campaigns with creative and prompt templates.

  • Guardrailed prompts leverage Anthropic- and Vertex-backed models.
  • Asset, telemetry, and token usage modules keep teams aligned.

TURF optimization service

Containerized Python TURF algorithms on Cloud Run Jobs to score each phase for reach, overlap, and incremental lift before spend is committed.

  • Jobs emit performance telemetry back to Looker dashboards.
  • Results feed automation scripts that adjust flighting in real time.
Outcomes

Proof that automation accelerates measurable value.

Darwin now powers Hoppr’s campaign pipeline with trusted data, guided workflows, and on-demand analytics-reducing waste while speeding experimentation.

Testing cadence

Teams can iterate on audiences multiple times per week.

99.95%Job reliability

Cloud Run Jobs auto-scale and alert before SLA risk.

Single sourceAudience truth

Segment governance shared across sales, ops, and agencies.

“DEVVIA translated our vision into a governed platform. Darwin’s assistant and TURF services make every campaign review data-backed and incredibly fast.”
Lena ParkVP of Product, Hoppr
Next transformation

Activate your own automation blueprint.

Devvia pods combine product leaders, engineers, and automation specialists that slot into your roadmap within two weeks - with the governance playbooks regulators expect.