akash@portfolio ~
› establishing secure channel...
AB
Akash Bishnoi
HomeAboutSkillsExperienceProjectsServicesFit CheckCertificatesBlogToolkitContact
All Projects
Marketing · Case Study

Pulse — Marketing Analytics

Unified analytics with attribution and forecasting.

Unifies paid, organic, and social into a single analytics surface — with first-party attribution modeling and campaign forecasting.

AnalyticsChartsForecasting

Product Demo

Interactive workspace

Explore the product workflow with a fixed local dataset. Controls and state changes are functional; the demo makes no external requests and requires no account or API key.

PulseMarketing performance
Data through May 20
AB

Executive overview

Revenue attribution and campaign efficiency

Attributed revenue
$301,200
+9.8% vs prior period
Media spend
$82,920
1.8M impressions
Blended ROAS
3.63x
Target 3.00x
Conversions
3.6K
+7.1% vs prior period

Attributed revenue

Last 30 days · all channels

Daily, normalized
Period startMidpointCurrent

Channel contribution

Position-based attribution

Campaign performance

6 active campaigns

CampaignSpendRevenueROASChange
Brand demand capturePaid Search$18,420$74,2804.03x+12.4%
Workflow launchPaid Social$22,100$63,8402.89x+8.1%
Lifecycle activationEmail$4,860$38,9208.01x+17.6%
Category educationOrganic$7,920$43,1605.45x+5.3%
High-intent retargetingPaid Social$12,840$34,7602.71x-3.2%
Non-brand acquisitionPaid Search$16,780$46,2402.76x+6.8%

Campaign drilldown

Paid Search

Brand demand capture
Campaign ID · BRAND
Active
936
Conversions
284K
Impressions
Cost / conversion
$20
Revenue change
+12.4%
Channel
Paid Search

The Problem

Marketing teams stitch dashboards across 6+ tools and still can't answer 'where did this revenue actually come from'.

The Solution

Pulse ingests channel data, builds a first-party identity graph, and runs a Markov-chain attribution model alongside a Prophet-based forecast — all in one cinematic dashboard.

Key Features

Channel Unification

Native connectors for Meta, Google, LinkedIn, TikTok, GA4, and more.

First-Party Identity

Server-side events and a deterministic+probabilistic identity graph.

Attribution

Markov-chain attribution with what-if simulation across last-N touchpoints.

Forecasting

Per-campaign Prophet forecasts with seasonality and confidence bands.

Goal Tracking

North-star metrics with anomaly alerts and Slack/email digests.

Cohort Explorer

Slice by source, geo, device, and product to find revenue lift hiding in plain sight.

Architecture

  1. 1Connector workers push to a Kafka topic; dbt models materialize into ClickHouse.
  2. 2Identity graph maintained in Postgres with edge weights from deterministic + probabilistic signals.
  3. 3Attribution and forecasting run as Python workers, results cached daily.
  4. 4Frontend uses Server Components for heavy slices and ECharts for live drilldowns.

Outcomes

12 hrs/wk
Reporting Time Saved
11
Channels Unified
8.2%
Forecast MAPE
+19%
Avg Campaign Lift

Highlights

  • First-party by default
  • Explainable attribution
  • Forecasts you can trust
  • Built for marketers
Discuss a Project Like This