Knowledge is Power: Lessons in Analytics from Expedia’s ‘fail-fast’ Culture
New York, NY (PRWEB) January 23, 2014 -- Knowing your customer better than the competition means more bang for your marketing dollar. It has put in place the necessary infrastructure and resources to be able to unearth consumer insights, and chart the customer journey by user segment. The aim being: to reach travellers with the right message at the right time. According to Ameya Karvir, director, analytics, Expedia, who is scheduled to speak at the forthcoming Smart Travel Analytics North America (http://eyefortravel.com/analytics) conference, slated to take place in New Yorker City (11-12 February), “analytics is embedded in the DNA of the business of Expedia. In fact, senior leadership is quite analytically driven.”
There is method in this approach. After all, data-driven marketing can lead to improved customer experiences and higher conversions as a result. However, while this looks great on paper, it requires an organisation to fully embrace analytics.
So what exactly does this mean? “Decisions, big and small, are consistently powered by sound analysis and data driven insights,” says Karvir. Expedia first focused its energy on marketing attribution and by 2009, “we had a flexible and scalable analytical platform that allowed our marketing organisations to test and simulate different attribution models to support marketing strategy.”
The need for speed
As online adoption continues to grow, the amount of digital information a user generates can be significant. “With the type of analytical tools and technology available in the market, the quality of data has improved quite significantly,” says Karvir, adding that this vast pool of data has become the foundation of initiatives such as marketing attribution. Complex patterns are continuously analysed to identify the ‘winning’ marketing effort. Such analytical and mathematical solutions are the force behind marketing investments and spend decisions that marketers typically take.
This is a fast-moving environment and so speedy delivery was a major focus for Expedia last year. According to Karvir, the company has embraced a ‘fail fast’ culture; when combined with big data analytics this has allowed the business teams to test different ideas and hypothesis at a much faster rate and incorporate the findings into operations quickly. What this has meant is the time taken to perform advanced analytics has fallen from a matter weeks and months to days, or even hours.
Rethinking attribution
There are three general approaches to attribution:
• Last (or first) click attribution
• Rules-based attribution (ie. Equally weighted)
• Algorithmic attribution (ie. Statistical)
According to Karvir, each approach for attribution needs a thorough understanding, not just of the methodology and its technical nuances, but more importantly how it relates to the specific industry in which your company operates. In a multi-channel, multi-product environment, user patterns vary across products.
First (or last) click attribution may be the simplest – and most popular, but it’s not necessarily the right one. Similarly, blindly using rule-based or statistical models may not really be necessary and could be overkill for a product or company. As Karvir points out, it is the responsibility of the analyst to understand the requirement of the business and ensure the model and approach aligns with the broader objectives and strategy of the organisation. It is also equally important to strike a balance between the level of effort required and the benefits the solution delivers. Assessing how analytics has helped to understand the key aspects of attribution is another necessary consideration. This means mapping out how conversions overlap across multiple online channels, while understanding how the combination of media influences latency or quicker conversion rates.
“It is quite evident that we live in a highly fluid, multi-marketing channel environment. Analytics helps us to understand that there is a considerable dependency of one marketing channel over the other,” says Karvir. For example, the changes made by paid search marketing could have an inverse effect on natural search. Such changes and decisions made in one channel can have significant impact on the economics of another.
Analytics also help firms understand that attribution is not entirely marketing driven; it has a product element to it as well. Some products have a longer shopping window than others and customers in the market for such products do behave differently.
Profiting from predictions
Marketers have also been working on getting predictions right to deliver greater profits from customers. Predictive analytics is about supporting change and rapid development while at the same time also mitigating a certain level of risk, explains Karvir. By continuously doing A/B testing and learning from those results, companies can introduce change incrementally.
Exploratory data mining is all about finding leakage, which is identifying pockets where something isn’t working. It could be that a bug is discovered, or that a customer struggle point is identified. It could simply be that a new opportunity is realised. All this can count in the final analysis, especially if this means connecting with customers better. By doing so, brands can increase the lifetime value of each customer.
Karvir sums it up: “My advice to travel marketers, and to all marketers for that matter, is that if you know your customer better than your competition, you have a better shot making the most effective use of your marketing dollars.”
Alesia Siuchykava, EyeforTravel, http://events.eyefortravel.com/smart-travel-analytics-north-america/, +1 (201) 204-1694, [email protected]
Share this article