Image by Robert Yang

What if we looked at our organization as a creative, design powerhouse on top of our core services and offerings?

What if every bit of relevant detail can be stored in inactive states ready to be deployed as creative ad units organized and automated into contextual elements by a homegrown application to position the publisher in a wide variety of angles?

What if these elements can be deployed and released into various content networks via a simple API connection?

A company’s marketing success can be determined by their ability to take a catalog feed and parse important elements like (product title, product group page and description lines) to build an application that transforms these elements into ad creatives, dispatched to a chosen platform through their API.

Having a proprietary in house ad automation tool for a channel helps decrease time spent in manual builds of campaigns, as well as ensuring all possible relevant combinations of ad units are available to blast networks which in turn drives traffic and sales for the organization.

This automation feature for marketing and communications can have wider applications.

We want our data feed product to be as flexible as possible. If we want to attach another datasource/feed that reflects our promotions calendar by product IDs, the application can pause and activate creatives to display promotions down to the product line.

The application can also determine which landing pages by product lines can determine the content of creatives. As the merchandising team works tirelessly to catalog product items this in turn perpetually refines the user experience for customers looking for the most relevant assortment of imagery, topics and knowledge.

When creative units are processed and generated the software attaches UTM parameters to the destination urls. For an eCommerce reatiler, running automated SEM programs, when items are out of stock, the application would create a flag and pause all relevant keywords in the ads system.

Building massive keyword sets (or in the case of custom audience targets) by combining queries with CTA’s and otehr qualifiers helps to multiply core interest targets tens of thousands of times.

For international campaigns, a catalog parser will also follow an index in which a language-ID matches a global product-ID to follow and track performance of all interest taxonomies regardless of what language it is translated in. This gives us flexibility to look at merchandising models on a global scale.

In order to optimize accounts with tens of thousands of target settings, it would be impossible to review every aspect one by one.

Building clustering schemas or taxonomies based on merchandising architecture we can retrieve sales data from our data warehouse and organize this info by thematic groupings to clarify which segments were under or over performing.

Creating logic to optimize cpcs that considered average position over time, quality score, match types, negative keyword traffic and seasonal shopping is tantamount to our success in driving top line revenue.

While the examples we've been observing above speaks to the core functionality of a program centered on pushing targeting elements through an API, this proprietary in-house tool can also operate bi-directionally and become a powerful data reporting and attribution center.

Reporting takes in real time sales, attributed to unique ad unit IDs (tied to accounts, regions, catalog properties) to help us make informed decisions.

In addition, we use the information to manage reporting of creatives / ad copy, in order to discover high performing assets across different product segments, promotional messaging and localized offerings for organizations with heavy focus in several regions.

We can sync daily activities to collect reports on a nightly basis. We can see how positions and quality scores fluctuate over time. We can also see quickly when average cpcs are hitting up against our max cpc (while positions have declined) indicating that bids have increased overall in the competitive sphere and more and more businesses are heating up the auction.

At I learned about tyhe inner workings of an API driven marketing environment that leveraged our catalog feeds particular to our business. .

Taking from the experience at, and joining Fandor, I worked on a viable product feature for marketing that would take our film library’s catalog feed of over 5,000 film titles into a similar parsing application to auto generate creative units based on xml feeds.

The parser connected to a few social media accounts like Twitter, Tumblr and Pinterest, serving up cover art and film descriptions as well as positioning CTA to auto populate status updates whenever a title appeared on the streaming service. With this app we were able to automate news feeds, thereby allowing the team to be more focused on tentpole promotions and acquisitions for highly acclaimed critical film collections.

This product feed for Fandor, which was optimized in perpetuity, was quite flexible in design. It allowed us to connect to other apis like content syndication platforms Outbrain and Taboola to expand our potential reach in a variety of networks.