Guest blogger on DMR: Kevin Hillstrom
It is my pleasure to welcome Kevin Hillstrom, President of MineThatData, on Data Mining Research. Kevin is a blogger at the MineThatData blog and he has accepted my invitation for a guest post on DMR. Here is his post:
Why Multichannel Forensics Matter
Kevin Hillstrom, President, MineThatData
Humans are good at detecting patterns. Mathematically, we’re not great at detecting patterns. But intuitively, we understand patterns, and can predict what the future might look like as a result of the patterns we detect.
We can look at the history of radio, noticing that listeners shifted from AM radio to FM radio. Then listeners began enjoying MP3s, creating their own playlists on their own players. Meanwhile, FM radio morphs into HD radio, Satellite Radio, and Internet Radio. As folks migrate from one form of technology to another, we intuitively view a future where music lovers have access to every song ever created. We imagine the impact this has on existing radio stations.
In business, we’re great at understanding the impact of competitors on our business model. We can easily see the impact of Google upon Yahoo! or Microsoft, theorizing what each brand must do to remain relevant.
Now take a moment to think about your own business. Have you introduced a new product line or a new service during the past two years? If you have, what impact did the new strategy have on your business? How did customers adapt to your new offering? Did customers who used to purchase existing products or services switch loyalty to your new offering? Did you find new customers to purchase your new offering? Or did your best customers increase spend by purchasing the new product offering and the existing product offering?
It is unlikely that your co-workers can answer these questions.
This is why I invented the concept called “Multichannel Forensics”.
Multichannel Forensics provide a framework for understanding how customers interact with advertising, products, brands and channels.
Pretend you are the CEO of Best Buy. You have an online channel, as well as more than 1,100 stores that sell merchandise to consumers. Could you answer the following questions:
- Does the website, bestbuy.com, add incremental customers to the business, or does the website help facilitate retail purchases?
- When a customer purchases a plasma or LCD television, does the customer continue to purchase other merchandise, or does the customer not visit the store or website anymore?
- Are loyal customers or new customers the ones most likely to purchase extended warranties?
- Is it better to open a new store, or spend the money it costs to build a new store on online marketing?
- Does Best Buy have loyal customers who purchase across all merchandise divisions, or one-time customers looking to meet a merchandise need?
Multichannel Forensics provide a framework for thinking about these topics. Products, brands or channels are evaluated on the basis of three dimensions.
- How loyal are customers who purchase from a product, brand or channel?
- How likely are customers to migrate to other products, brands or channels?
- Which product, brand or channel is responsible for acquiring the most new customers?
The framework helps the CEO understand the role of every product, brand or channel. In our Best Buy example, Multichannel Forensics might indicate that the website is the most cost-effective way to acquire new customers who eventually become loyal store customers. Or Best Buy might learn that appliance customers are unlikely to purchase electronics. Obviously, I don’t know the answers to these questions, but it is important to understand how to resolve these issues.
Multichannel Forensics allow the data mining analyst to build simulations that illustrate the best way to grow a brand. The methodology allows data mining analysts to have a relationship with the executive team. Executives have very little actionable customer information. By helping the executive understand the best way to grow a brand, data mining analysts provide a valuable service that is sorely missing from most analytical discussions these days.
Helping CEO’s understand the relationship between customers, advertising,
products, brands, and channels.