Case Study: Three Surprising Results from A Contentgine Data Science Analysis

Date

When Ben Luck, chief data scientist for Contentgine, began analyzing the performance of the company’s client target account lists (TAL’s) to seek better ways to improve the clients’ targeting efforts, he noticed things that he immediately felt he could help correct.

Luck was a prime architect in the company’s trademarked methodology for producing audience engagement with content, called the Perpetual Engine®.

That methodology was based on the company sending content to over 50 million emails every month and capturing what content they engaged with. The data from those engagements could be applied to the company’s client TAL’s and significant improvements might be made.

Luck wanted to tackle the challenge of continuous improvement in his clients’ TAL’s because it is an area that requires every marketing department to regularly visit and analyze who they believe is the proper audience for their offerings. Marketers have told the company that they believe they have, in the past, likely wasted large amounts of demand generation spend on poorly targeted marketing campaigns.

The misalignment can come in several ways: not properly identifying the right business segment, not offering the right type of content, not properly targeting the right persona or role in a company, or even not properly titling business cases and other content to drive interest.

“I’ll just say that a target account list always has to be scrutinized for performance and accuracy,” said Becky Carr, Chief Marketing Officer of Tangoe.

“Marketers are always looking for ways to prove their target lists are and will remain, effective. The process hasn’t changed much over the years: you run campaigns, you test a proven message with different TAL’s, and note who responds to the message. It’s iterative and time-consuming, and any improvement or augment to a TAL is a welcome development.”

Luck set out to analyze the data that Contentgine itself generates through its own marketing outreach, which is a critical function of the company’s process. In order to generate an audience for its clients, Contentgine utilizes a large and complex process of sending content to interested parties, a process it calls the Perpetual Engine.

 

Data analysis that informs better client TAL’s

The perpetual engine dataset

As previously described, Contentgine’s monetized assets are the lists the company creates through its internal marketing process it calls the Perpetual Engine. This process involves sending over 50 million emails, newsletters, or microsite web links to interested audiences monthly. Those emails contain case studies or other marketing materials created by over 21,000 companies with 400 curated solution sets, in a collection of marketing content that has reached over 500,000 content assets.

When the target persona reads or downloads a particular case study or other content, Contentgine captures that persona’s identity, wraps first-person information about that persona into a list (company, role, etc.) and supplies those clients in that business segment with those first-person lead lists.

By doing this literally every day to keep audiences engaged and informed, the company has built its own vast dataset of what personas interact with what kinds of content. That dataset can be analyzed with some precision, and data scientists like Luck can examine a customer’s TAL versus a similar TAL that the Perpetual Engine has used successfully.

That allows Contentgine to make informed suggestions to the clients on better ways to effect content engagement. When Luck did this analysis for three companies in his investigation, he found the results to be both impressive and a little surprising.

 

Deconstructing data to learn what works

Luck started his investigation by selecting three different clients and then supplementing their targeting with data he had derived from the Perpetual Engine dataset. For the investigation, he chose a cloud communication company, a financial ERP company, and a digital signature company.

Luck began by deconstructing the data into the target elements that go into a TAL:

  • IndustriesandBusinessSegments – CompanySize
  • SpecificTargetCompanies
  • EmployeeTitleorRole

And other TAL elements. Luck then matched Contentgine’s dataset for marketing to the same type of client and examined whether Contentgine had a higher engagement rate with the same audience that each of the three companies had targeted. He then set about looking at the assets the three companies were offering up to their potential audiences: what type of asset (case study, whitepaper, video, etc.)? What marketing messages were these companies relying on, both in an email or newsletter, plus in the name of the case study or other asset. He then compared those messages to messages and titles that were most effective in the Contentgine dataset.

The results, Luck stated flatly, were very surprising.

 

Data-informed TAL’s

Effective supplements to target marketing

Luck’s next step was to create his own TAL for each target customer to supplement those companies’ own TAL lists. Luck ran campaigns with his own TAL’s and with the client TAL’s and then gauged what and who were more likely to engage with the content. His conclusions were as follows:

 

CONCLUSION 1

Contentgine TAL’S Improved Engagement 10-20x

Every one of the three companies showed a significant improvement in the TAL engagement rate, with a 10-20x improvement depending on the company and the asset. By building the Contentgine TAL from the company’s internal dataset and combining that with the elements in the client’s TAL that were shown to be effective, Luck had landed a vast improvement method that required only a little extra analysis work.

“I guess I’m not surprised at the improvement rate, when you consider how much data we have internally for content engagement,” Luck said.

“We send over 50 million content emails monthly, and process over 1 billion intent signals every quarter, so there is likely no company anywhere that has more data about content engagement. It took a little data science magic for me to extract the right information from that dataset, but I’m very pleased with the results, and I know our clients are also pleased.”

“Even a 10x improvement in TAL engagement is of huge value to a marketing department,” Becky Carr added, “and a 20x improvement is a dream. There hasn’t been much movement in the intent and content marketing arenas that focused on TAL improvement, so to me, this is a significant breakthrough.”

 

CONCLUSION 2

Client Misalignment on Business Segments Can Be Improved

Luck’s next revelation was a mild surprise that he wasn’t necessarily anticipating, but one that stood out, nonetheless. In creating his own TAL’s from the Contentgine dataset, Luck noticed that content assets similar to the clients’ assets were performing well in business segments the company did not originally target.

This misalignment of business segments could result in a company wasting a large marketing budget in offering products or services to a segment that it believes is a target, but is, in fact, either an under-performing segment, or one that has adjacent segments that the company didn’t consider.

“One client in particular was a case of missed market opportunities. In this scenario, the company firmly believed that its offering was a solution primarily for small business and aimed almost all of its marketing there. We were able to demonstrate to them that similar assets to theirs actually performed well in the mid-sized and large markets. Needless to say, that was a surprise to them.”

“Aligning your marketing outreach on the right market segment is both an art and a science, and misalignment happens to everyone at some point,” Becky Carr stated. “Analyzing your existing customer base can get you only so far, so you have to spend time and money testing markets and doing so in a large enough way that you can draw conclusions. By having Contentgine actually suggest new market segments that the client company might have missed is a huge advantage to the company’s offerings. You are literally saving that client money in testing and analysis.”

 

CONCLUSION 3

Wording and Asset Titles Actually Matter in Engagement

The most surprising conclusion that Luck reached was the one that required the most investigation, and one that was subjective to his interpretation. But he believes he can conclusively suggest that the way that a client words its engagement messaging in emails and other outbound marketing, and the actual titles of the content assets being offered by that client, might make the difference in whether a client clicks on the asset to read it.

Luck analyzed a number of email interactions with a client base and noticed that the more specific the title of the content piece, the less likely the target persona was to open and read it. Luck saw that the company’s logo was a differentiator – the more well-known brands were better trafficked – and that the background information on the asset might lead to a clickthrough on the offered link.

But Luck also noticed that there was a drop-off in certain cases when the actual content asset was presented. After reading enough content asset titles to draw a conclusion, he kept it simple: don’t be too specific with a content asset title.

“This surprised me because it’s something none of us had ever even considered before,” Becky Carr stated.

“Who ever thought a case study or whitepaper title might be a repellent to being read? It’s just good data science to be able to show this and it’s now something we marketing folks can test to see how titles affect engagement.”

But Luck also noticed that there was a drop-off in certain cases when the actual content asset was presented. After reading enough content asset titles to draw a conclusion, he kept it simple: don’t be too specific with a content asset title.

 

Luck’s study is just the beginning…

“I believe my value to our clients is that I relentlessly analyze their data versus our data and can make data-driven, informed changes to their content outreach efforts. This is just the first of my investigations and was done with a small sample set. My next effort will be even more analysis on even more client TAL’s and building an automated way to do that at scale.”

“When your investigations improve your clients’ performance,” Luck concluded, “it’s as much as you can hope for as a data scientist.”