Author: Jessica Hawthorne-Castro
Original Link: Datanami
Date Published: June 1, 2017
Walmart is a global leader in using data analytics to form insights. For example, the company previously used data generated during hurricanesto better estimate demand for emergency supplies, so it could be best prepared to meet customer needs. The company focuses on both reactive and proactive analytics, so they can thoroughly review past problems and spot upcoming errors in advance.
Enterprise companies such as Walmart that are successfully using Big Data has some commonalities. They have the technical expertise and the right systems in place to manage massive sets of data and know how to learn actionable insights from the data. For marketers that want to improve their results through data analysis, it can be tricky to correlate what the data is saying and how that could adjust marketing strategies.
Here’s five best practices companies can employ to refine their campaign performance through big data:
1. Leverage Pixel Tracking
Marketers should not overlook their company’s websites as data sources that provide them with a wealth of information about customer behaviors. A primary method of capturing website data is through pixel tracking, which can be placed on mobile sites, landing pages, and any other web-based channel. Marketers can also track social media with pixel tracking to better understand Facebook ad performance and the visitation patterns of followers. Pixel tracking not only provides information on what visitors click and how long they stay on certain areas, it can also provide device data. Marketers can then tailor campaigns, by for example matching certain product lines with mobile customers.
The broader point of having this trove of data is to create “profiles and personas” of groups of customers. Armed with such profiles, marketers can then correlate behaviors with buying decisions or indecisions. For example, they might see a longer “mouse hovering time” with certain products, and could then propose pricing adjustments or added incentives to make those products more enticing.
2. Use the Cleanest Data
Big data is only as good as the underlying data – it’s not a magical technology that can fix sloppy data. If a company pushes poorly organized data into the analytics solution, then the resulting insights will be unreliable even if they are pulled by an expert data scientists. Marketers should demand that data going into the system will be scrubbed and organized. This does not mean data streams should be restricted, they should just be well constructed and reliable, and so the end results are meaningful and have real value.
Proactivity is crucial, and companies should put in the work on the front end to improve their data quality instead of attempting to make changes during active campaigns.
3. Employ Statistical Modeling to Predict Future Results
TV campaigns can be improved by using data collection to look at past results, and then making the right tweaks for the future. Marketers must have access to metrics on demographics, second-screening, airing size, and Nielsen weighting points in order to build statistical models.
Clean and granular data is essential in order to inform regression analysis, logistic regression, and cluster analysis to dive deeper into campaign performance. TV campaigns do not have to be unmeasured branding boosters. With analytics, agencies and marketers can show measurable results from TV.
4. Improve Demographics Targeting
Rich big data is ideally suited for targeting. It gives marketers the info they need on the location of customers, the devices they are using to access content, their behaviors while on social sites, and even their search habits. Pooled together, marketers can develop correlations, such as relationships between devices and products, or location and preference for mobile or landing page. Marketers must have the right tools in place that allow them to adjust future campaigns based on the improved targeting, so they can present adjusted granular messaging.
5. The Benefits of Mixed-Media Modeling
Mixed media modeling entails careful analysis of both sales and response data. Marketers can then review various channels and eliminate laggards while pushing more budget towards the top performers. This process will improve over time, as the data becomes cleaner and more structured, and marketers and data scientists will improve their analytical talents. Such modeling is especially important in the fragmented media landscape, as there are many choices for marketers and they need to narrow down the best channel prospects.
Using big data analytics to improve ROI is a necessity for today’s modern multi-channel business. Success with big data takes a measured approach, where marketers are focused on pairing insights with past and future actions. Big data provides marketing with a much-needed measurement tool that can use to prove their result and improve their campaigns year-over-year.
About the author: Jessica Hawthorne-Castro is the CEO of Hawthorne, an award winning technology-based advertising agency specializing in analytics and accountable brand campaigns for over 30-years. Hawthorne has a legacy of ad industry leadership by being a visionary in combining the art of right-brain creativity with the science of left-brain data analytics and neuroscience. Jessica’s role principally involves fostering long-standing client relationships with the company’s expansive base of Fortune 500 brands to develop highly strategic and measurable advertising campaigns, designed to ignite immediate consumer response.
