We are not all the scientific data – but that does not mean that we should not roll up our sleeves and gain the full potential of our Big data and marketing initiatives focused on data. My colleague, John Kucera – scientific data itself – explained in his recent Adobe Summit session that all of us are born with a natural curiosity for science. However, in the words of Carl Sagan, “Every child begins as a scientist born, and then we beat them.” That means it’s within our reach to think, act and behave as data scientists. As traders, we just need to commit to retrieve our scientific advantage – to act, think, and testing as scientific :. data scientific
curious and Curiouser
all starts with finding our innate curiosity and skepticism: two keys that separate scientific “- and scientific data” – processes the rest of the pack of passionate curiosity should. manifest as a relentless pursuit of the truth in each of your marketing campaigns Suppose you change your banner for winter coats, causing conversions go through the roof -.! incredible But, that does not mean you should always have winter coats hero in this place.
There is curiosity – what was about this creative in this investment that led this success? And there is also skepticism – was Simply brilliant thought that made this magic conversion occur? You must put your scientific hat, fire up the Bunsen burners, and work to discover this contributed to this victory, so that you can replicate the experience again and again.
data science is an interesting intersection – in the end, where the marketing and science meet – an area in which we are all razor focused on solving problems. That’s what these two areas have in common, and that is where marketing can dive and put their skills, analytical thinking, and machine learning Sidekicks to great use.
feel even stuck on the way? According to John, he began by exploiting your curiosity, asking questions, exploring data and assumptions. Depending on your business and industry, some good questions to get you started could be:
- Which customers should I focus on, and which the public can I influence ?
- Where are those customers come? SEO? Mobile? Web? At the store?
- What changes can I make to influence their behavior?
- Where do I change? My website? mobile experience? App? At the store?
- how key performance indicator (KPI) improvements that I can try?
The ideas you find exploring – usually by means of a platform such as Adobe Analytics – lead you to ideas to create hypotheses. Need some examples?
- The students prefer hotel deals in which they receive a free night after staying three nights.
- The people of the western United States prefer skiing to snowboarding.
- visitors respond to ads on the left side of the page above the banner.
Again, these are assumptions based on what you have observed through your data mining. There seems to be significant trends, but we must better analyze and interpret data to see if this is really the case. To do this, we must determine whether it is causal or correlated. Because correlation is not causation . Engage the memory -. Maybe even jot down and pin it on your wall
John shared an amazing example of this at its summit session. Examining historical data, it was noted that, as temperatures around the world have increased the number of pirates sailing the seven seas has decreased dramatically. Obviously, this dip in the global population of pirates is causing global warming. Let me repeat :. Correlation is not causation
Acting as a scientist
The correlation is not causation. healthy skepticism – and yes sense too – is essential to the process of science. But traders must also draw in other key traits for success in the face of scientific data -. Observers and casual fans, even as simple
1. Precision Awareness
Each measurement has an amount of error associated with it. Your goal? Avoid overcommitting when your upside potential is overshadowed by the error associated with it. So if committing much resources and budget will give an improvement of two percent that comes with a margin of error of five percent – do not . This two percent jump is just noise. But if the error measure is .05 percent, and your head is two percent, shout it from the rooftops and go !
With some skepticism, of course, on scenarios that are so very positive.
So how do we get the error measures? An example is the Adobe sample size calculator target, allowing you to input data about visits to your site, conversion rates, and how long you want to run a test. Then it spits out an overview of how many items you can test and what level of error you will get as a result. As simple as that.
2. Establish your Control
To excel in this field, you need the mental discipline of scientific – or something pretty close. Back to the example of winter coat: let’s say you run this promotion in the late fall when people really need jackets. You had a prominent banner on the home page boasting a amazing coats brand, and have seen tremendous conversion rate.
But you are a scientist. And you are a skeptic.
Without control, it is difficult to say how significant that was promoting. But with control – perhaps a group that did not show the coat promotions – you’ll be able to begin to analyze what is actually happening. Perhaps the promo a make all the difference, and people who are not exposed did not convert nearly as well. Or, perhaps, people come to buy coats -. It was November, after all
3. And (new) – SKEPTICISM
Ironically, if you are quite skeptical data and how it was collected you gain greater confidence data as it passes your tests. This allows you to trust the results against-intuitive and use them to your advantage. Once I see that this promotional led these results – thanks to have control, establish a margin of error, and to test other options – I am convinced that this campaign moved the needle, and I can begin to deploy additional elements confident Because when you do -. ultimately -. trust the data, you can reap the benefits in a big way
Let Technology Be Your Data-science Surrogate (But take Yourself Credit)
Again, you do not need not a scientific data – after all, you are a marketer, aren ‘t you? But as we discussed, this does not mean you can not use data science as a better marketing. Adobe Marketing Cloud offers built-in data science, including predictive analysis, the discovery of the audience, and of course, Adobe target statistics driven decision engine that powers the personalization of website such as recommendations. Together and individually, these solutions take the heavy lifting and many tedious parts about testing your assumptions, watching thousands of correlations instantly based on historical data and real-time. By automatically customizing the content to visitors, these systems can improve conversions or revenue with learning machine who finds correlations between the properties and responses visitors.
Above all, trust your autopilot. Yes, skepticism is important, but it is also important to let your machine partner optimize and customize – without your backseat driving. This might be easier said than done, depending on how you approached the customization in the past, but it is worth it.
Your next step? Focus on reinvigorating your natural curiosity so that you’re not just ask why something is happening, but also unleash your inner scientist to explore, hypothesize, test and optimize. Be smart and be scientific in your approach – one powered by precision, with emphasis on control, and of course, healthy skepticism. Automation can take some of the heavy lifting and doubt on the process, provided that you can trust your sidekick.
When the pieces come together, it is powerful – and you do not even need to be a scientific data to achieve serious end results. Just roll up your sleeves and tap into your inner science -. It is definitely kicking around there, waiting to be released on your next marketing initiative