Tips & tricks to develop a ‘data’ culture in agencies
As a performance agency, iProspect has always been home to data-savvy advertising practitioners, handling paid search accounts with millions of keywords requires a certain comfort level with data manipulation on a large scale.
However, we are rapidly approaching the point where data that teams are handling is going beyond what desktop computers, macros, pivot tables and VLOOKUP in Excel can adequately process.
In addition, clients are demanding more from their agencies, and it is often data capabilities that will make the difference.
For example, we may need to work out the diminishing returns of further investment in a channel or campaign, model the optimum mix of channels to hit a required volume of sales of leads (one million seems to be a popular target that everyone aims for), attribute value to every consumer touch point in an online sale, or work out a predictive model that takes external data signals into account.
This kind of advanced data wrangling requires a combination of tech skills to work with APIs on data ingestion and advanced math skills, which are both typically beyond the normal capabilities we see from candidates gravitating to work in advertising and media.
These requirements will only intensify as media is further traded and measured digitally.
In the past, there has been a mix of motivations that have driven people into advertising and media as a career, but as an industry, we haven’t done a great job of attracting data talent from potentially more lucrative careers in finance. If you’re an aspiring data scientist, there’s a reasonable chance that you have never even considered media and advertising as a career option. However, this industry has some really interesting problems that could certainly be of interest to some of the prospects who are keen to avoid a buttoned-down work environment.
To try and address this issue, last year Google and the IDA teamed up to launch the Google Squared Data programme which was developed to specifically promote our industry to recent university graduates, who have studied mathematics and computer sciences. After an intense couple of months studying at Google (with a lot of external speakers from the industry), the interns were then released into the wild to work with agencies for another eight months. The programme was, and continues to be, successful for us, not only as an agency, but also for the industry.
That said, we need to be doing more, and not solely relying on Google to do the hard work for us. Agencies need to adopt a data culture to attract, nurture and develop data talent.
Some ideas:
#1. Create an organisational structure that encourages data-driven thinking with an environment of ideas and collaboration across the agency
#2. Unite the different pockets of expertise under an informal data community that are supported by management with an attitude of experimentation, exploration and some risks
#3. Fund cloud infrastructure that is not directly linked to a specific client outcome to enable experimentation
#4. Promote a culture of curiosity – asking questions and answers based on facts and data
#5. Build academic outreach programmes to enable university students to work on data problems as part of their course work (gaining exposure to the industry)
#6. Take on interns and graduates with data skills but do notbury them in mundane reporting. Instead, empower them to work on automation and visualisation to shift the task balance to analysis and insights rather than mechanical reporting
#7. Encourage organisation-wide data skills with data or programming boot camps – there may be more interest that you expect
#8. Sign up for one of the many online data science or introduction to programming courses – they are low-cost and will open a window to the future of how agencies will be operating.
To conclude, here is an example of how experimentation with data can drive business results.
Curious about the new breed of cloud services that are backed by machine learning algorithms, one young member of the team experimented with a lead-generation performance campaign to find out if there was anything in the data that could enable further optimisation. It turns out that there was a very strong correlation between owners of luxury SUVs (a data point captured in the lead form) and leads turning into confirmed sales. Armed with this insight from the data, the sales team was able to concentrate their calls on leads who have a luxury SUV. This one insight resulted in a 180 per cent increase in sales for exactly the same budget. Additionally, it opens the marketing programme up to a whole new avenue of audience targeting that converts at a higher rate.
The post Tips & tricks to develop a ‘data’ culture in agencies appeared first on Digital Market Asia.
