Question: I've been practicing web analytics for a decade. I'm perfectly happy analyzing conversion rates and executing multivariate tests, and my boss has never once asked me for something like this. Why would I ever need to execute an Online Marketing Simulation?
Answer: Your business leaders are telling me they want you to do this type of work for them. Your CMOs, CFOs, and CEOs communicate to me that they want to know what will happen to their business in the future. Your leaders are being pressured to understand if their business is capable of significant growth in the future. They want to know which micro-channel is best positioned to provide future growth (e-mail, affiliates, search, social media). They want to know if e-mail marketing performance improves if they increase spend in paid search. They want to know if making radical adjustments to offline advertising will impact sales online. They want to know what happens to the rest of the business if they decide to end all affiliate marketing programs. They want to know what total online sales will look like in 2014. They want you to give them the answers to these questions. I think your company is best served when you can provide "the C-Suite" with these answers.
Question: All of the Web Analytics bloggers and experts seem to be ignoring this topic. Is it possible that they know what's best for us and that you are talking about something that just isn't part of the typical web analytics tool kit?
Answer: Of course it is possible that they are right! But let's be honest. The Web Analytics community needs to integrate all offline and online data, and then become proficient not only at measuring what is happening, but forecasting what it means to the future health of a business. The vast majority of Web Analytics bloggers and consultant experts haven't had a need to solve a problem like this, especially when e-commerce was growing organically at 20% or 30% a year. Those days are over. Few e-commerce experts and web analytics experts have had to deal with the pressures of a stalled business, so they haven't ever had the need for a tool like the OMS. Over the next five years, you will see an evolutionary push toward advanced tools like the OMS. The marketplace and a stagnant economy will force it to happen.
Question: Your presentations are too vague. I don't think you're doing a good job of communicating this topic to your audience. What can you do to improve your presentation of this topic?
Answer: I'll try to improve and offer more concrete examples. This isn't an easy topic to communicate. I'll continue to work hard to improve the communication of this topic.
Question: Google gives away Google Analytics for free, so why shouldn't you give your programming code away for free? If you believe so strongly in this concept, why not evangelize it by giving it away, and then find some other way to monetize your activities?
Answer: I invested more than half of my professional career writing and testing the programming code necessary to create an application that predicts sales by micro-channel. This tool is perfect for the forward-looking CEO, an individual who wants to understand the future of an online, retail, or direct marketing business. I've invested tens of thousands of hours on this topic. That being said, I did outline the exact methodology in the presentation. If you are handy with programming code, and are a wiley analyst, you can replicate the algorithm based on the information in the presentation. Go do it!!
Question: "Our web analytics data is cookie-based, and we all know about the problems with cookies. So I think that invalidates the Online Marketing Simulation, right?"
Answer: "Absolutely not. Use the data you have, and where you don't have good visitation data, use your purchase data, which is much more important anyway. Please do not become paralyzed because you have incomplete data!"
Question: Have Omniture, WebTrends, Unica, or Coremetrics approached you about incorporating your algorithm into their software applications? It would be great if we could do this right from the platform we do all of our analyses from. Why not partner with them if you want this tool to become widely used?
Answer: No, they have not approached me about incorporating the methodology in their software applications. But representatives from every one of those companies are reading this series and are actively downloading the presentation, based on my blog/site web analytics data. The same holds for the major Web Analytics bloggers and consultants, my web analytics data show that they're all downloading and reviewing the content. If you like what you've been reading, communicate your interest to them, because people are paying attention.
Question: Must we use a Factor Analysis as part of the simulation routine? Must we use regression models to rank customers from best to worst? Why make this so complicated?
Answer: Absolutely not. Create your own segmentation scheme, and run the simulations with your version of segmentation. It's the simulation/algorithm that matters, not the methodology that is used to place customers within segments.
Question: How is this different than the Multichannel Forensics work you've done for catalogers?
Answer: The core of the Multichannel Forensics methodology is the core of the OMS tool. We're simply shifting the focus now, from figuring out how catalog advertising impacts the business, to figuring out how a myriad of online micro-channels and merchandise categories interact to impact the future of your business. Think of this as Multichannel Forensics on steroids. All of the concepts in the Multichannel Forensics book hold up well in the OMS framework.
Question: What is the most interesting thing you've learned while running Online Marketing Simulations for clients?
Answer: The way that subtle changes in the merchandising strategies of advertising channels impact the future merchandise preferences of customers. When you optimize e-mail campaigns based on the merchandise sold in e-mail campaigns, you impact the merchandise divisions that will sell well in the future. When you focus on various keywords, you impact the merchandise divisions that will sell well in the future. When you focus on discounts and promotions, you dramatically alter the responsiveness of your customer file to full-price merchandise in the future. It is amazing how things like discounts and promotions act like a virus (or a poison) that spreads through the whole customer ecosystem, reducing your ability to sell full-price merchandise. The OMS is perfectly suited for illustrating the poison of discounts and promotions, showing you how a heavy season of promotions in Fall 2008 dampen full-price business in 2011. The OMS tool literally allows you to see how you "infect" your business with viruses, even though you strongly believe you are optimizing your business for positive results today.
Give the presentation a read, and let me know what questions you have. If we get enough questions, I'll dedicate another post on the topic.
I'm reading your MultiChannel Forensics book and it has helped me understand your presentation, as well as realize how far I have to go. More comments to follow as I get up to speed.
ReplyDelete