GroupM delivers vital client insights with Fivetran

“Our responsibility is onboarding data that gives value back to our major clients by helping them decide when and where to spend their advertising budget. Fivetran helps us to do that.”
— Herman Mull, Data Analyst, GroupM

The Situation:

Based in Oslo, GroupM is a global media agency group and the Nordic region's largest media agency group, serving over 200 clients with technology-enabled services that connect with premium partners, including Google, Facebook, and Amazon. GroupM provides a shared services-type function for the other agencies in the group – MediaCom, Mindshare, Wavemaker, MediaPlus – and is responsible for collecting data and measuring the effectiveness of their award-winning advertising campaigns.

The Challenge:

Obstacles to Accurate and Timely Data  

A data team of five was using Supermetrics to pull marketing data directly into Google Sheets, but limitations were becoming apparent. Pipelines would occasionally fail, caused by issues that were hard to detect, never mind resolve, and there were formatting problems with the spreadsheets as well as manual errors.

Preparing data for analysis in Google BigQuery, GroupM’s data warehouse was proving to be labour-intensive work – at a time when clients want faster access to more insights.

“It was not really a scalable solution,” recalls Herman Mull, Data Analyst at GroupM. “We needed to spend less time manually building connections and maintaining them; we needed to be able to push data directly to BigQuery without going through Google Sheets.”

One particular Norwegian client, with a broad business portfolio spanning retail and hotels, was looking for dashboards that could handle historical data analysis as well as day-to-day reports. GroupM was determined to step up to the mark with a more robust solution.

“We could do it with what we had, but it took a lot of time and we had to constantly monitor it for errors. It was a struggle for such a small team,” he says.

The Solution:

Streamlining Data Processes

Fivetran was identified as the pipeline solution, providing connectors to pull data from a wide variety of data sources – including Facebook and Google Ads – directly into Google BigQuery. The GroupM team was impressed by Fivetran’s self-healing and zero maintenance architecture, along with fast access to pre-configured connectors that could speed up ingestion from up to 15 main sources.

The goal was to gather all the data in BigQuery to build automatic reporting. Every client has different marketing KPIs (Key Performance Indicators), but metrics for impressions, clicks and conversions were always in demand. Setting up a connector is now easy, according to Herman Mull, because all the documentation is directly in front of you.

“For a Facebook connector, for example, if I want to add a link to ‘clicks’, which is one of the main KPIs most of the clients match, the documentation tells me what it is and what to look for,” he says.

Part of the plan for the next phase of development is to use Connect Card, an embeddable Fivetran setup feature, which authenticates user credentials via a secure HTTPS connection and avoids any work on the company servers. The Connect Card popup would allow GroupM customers to input their login credentials for the data source directly and initiate the sync, avoiding the need to share proprietary data with GroupM.

• Integrated instructions and set up tests would walk customers through any required configuration
• In case of a connection failure, actionable feedback to troubleshoot would be provided without users needing to reach out directly to the GroupM
• Connect card is easy to embed with only two calls to Fivetran’s REST API and a redirect URL to display the setup form

The Outcome:

Forging Closer Client Relationships

The main reason for changing to Fivetran was to standardise processes, accelerate time to insights, and to free up the data team from having to manage and maintain pipelines.

“Typically it would be an API not responding or the format being wrong and preventing it from feeding into the query,” explains Herman Mull. “That’s all taken care of with Fivetran.”

The time saved equates to three to five hours per month, per client. GroupM has 15 key clients they provide marketing analytics for, so in a busy period Fivetran saves as much as 75 hours a month, or nearly two weeks of work. Fivetran also backfills GroupM customers’ historical data, saving the team further time and avoiding time-consuming manual syncs.

Reports are presented to the client via dashboards that are updated daily.

“Our responsibility is onboarding data that gives value back to our major clients by helping them decide when and where to spend their advertising budget. Fivetran helps us to do that,” Mull says. “They can basically follow our clients' campaigns across all marketing platforms. The big win for us is that all sources are gathered into one place using Google Cloud for Storage and Datastudio for visualization which would normally be viewed manually in different reports across different systems.”  

Going forward, the plan is to move beyond marketing and use the data for sales modelling, leveraging AI and machine learning for statistical analysis of the factors that influence a purchase. Increasingly sophisticated analytics are a direct result of using Fivetran.

“Instead of spending our time maintaining pipelines and collecting data, we are actually using the data,” says Mull. “That's basically what our goal is.”

Visit our Powered By Fivetran Resource Center to learn more about the benefits of a modern data stack for your agency.

[CTA_MODULE]