Inventale

Our Clients and Partners

Case 1

Improved Delivery for a Network with High STR

Improved delivery in high STR environment

Challenge

Major online cinema (>1 bln imps monthly) was constantly losing part of their revenue due to under-delivery issues resulting from over-selling

Details

Custom in-house Ad Server without API support + 3rd party RTB-based Ad Server as the 1st priority for selected premium sales
High STR but many unfulfilled contracts (35-65% of under-delivered orders monthly)
High dependence on user traffic impacted by calendar and uncontrollable events (from week-ends and holidays to weather conditions)

Solution Specifics

Inventale forecasting does custom Ad Server delivery logic simulation to provide accurate forecasting levels
Integration with the customer Ad Server via logs (as there was no API)
Support of RTB inventory as the 1st priority in forecasting (higher unpredictability)
Historical trends analysis to detect changes in audience behavior

Results

Monthly campaign delivery increase +25-30%
Retention rate increase (due to better contract obligations fulfillment) +17%
ROI increase + 14% (monthly average)
Case 2

Improved Sales of Audience for Narrowly Targeted Campaigns

Improved Sales of Audience for Narrowly Targeted Campaigns

Challenge

An industry-leading publisher (>1.5 bln imps per month) with substantial TV content had issues with delivery of narrowly targeted ad campaigns due to cannibalization of high-value inventory

Details

Media planning in terms of unique users rather than in terms of available inventory permitting TV advertiser to directly address the target audience
A combination of video and media inventory, with different inventory origins (the web, mobile, and SmartTV) and different user behavior for each category created forecasting complexities
Highly variable inventory dynamics due to content nature (different TV series, launch of new seasons, etc.)
A large number of contracts with extremely narrow targeting parameters resulting in many under-delivered campaigns even when overall inventory availability was rather high Last-minute
Inventory purchase from partner networks when lack of required inventory and under-delivering campaigns were in place

Solution Specifics

Analysis of inventory packages to avoid cannibalization and give recommendations on pricing of different inventory pieces
Specific events behavior simulation (i.e. TV premiere, changes in user interests as TV series unfold, etc)
Automatic analysis and adjustments of inventory spikes to smooth their influence on forecast results
Additional heuristics to model video ads specifics
Automatic alerts to forecast and address early stage under-delivery risks

Results

STR uplift from 57% to 84% (monthly average) of narrow-targeted campaigns
14% total revenue uplift due to sell-out increase and solved on-time delivery issue
The number of campaign prolongation cases decreased by 64% (improved delivery of narrow targeted campaigns)
Case 3

White Label Forecasting Solution for Premium Sales Platform

White Label Forecasting Solution for Premium Sales Platform

Challenge

Major media holding corporation (>10 bln imps, >50 sites) developed a platform for advertisers and agencies to book and run ad campaigns directly through the platform UI

Details

Stable high-quality forecast by any slice of inventory with results immediately available to advertisers
Quick response to each request (< 10 sec per any complex forecast request)
Multiple simultaneous forecast requests (up to 100 concurrent users accessing the system)
Availability of service - 24/7
Integration with CRM to incorporate financial information on rate cards and inventory pricing

Solution Specifics

Custom built-in Inventale forecasting solution through custom API's and integration with customer's CRM
Hadoop cluster for parsing and forecasting load balancer to process high load
Constant monitoring and internal testing to guarantee performance and quality as network changes over time

Results

3-8 sec per request, 10000+ requests daily
Balance between forecast accuracy and response speed determined by the client
Forecast level accuracy 85%+, even for the most complex targeting requests
STR increase + 23%, retention rate increase +31%, ROI increase + 16%