eCommerce Analytics

There’s really not much to say about the tight end predictions except to repeat what I said for both the running back and wide receiver predictions. That is, the yardage totals and receptions are depressed. However, when taken in context, the rankings make quite a bit of sense. Well, if you expected Dallas Clark to be the top-ranked tight end. Click for the list.

Read Tight End Predictions for the 2008 Fantasy Football Season »

As I promised yesterday when I gave you our running back picks based on our predictive models, today is wide receiver day. And just like our running backs, you’ll find that the yardage totals of our wide receivers are somewhat depressed. That is, one might believe that some receiver will break out for 1300 yards or more in the season. And it’s likely someone will. But my model is very conservative.

Read Wide Receiver Predictions for the 2008 Fantasy Football Season »

Well, I apologize for not getting these out sooner but my draft just finished this week so there was no gun to my head in getting these models run. But in case you’re still curious about some regression-based predictions for this fantasy football season, I’ll drop the running backs on you today and then come back with wide receivers tomorrow and tight ends on friday.

Some quick notes on the running backs. In general, you’ll notice that the rushing yardage is quite depressed. This is largely because running backs tend toward entropy. That is, once they have a big season, the next season’s rushing total doesn’t quite live up to that previous season. Call it what you will: overwork from the prior year, increased injury likelihood, or teams keying on that back. Whatever the case, you are seeing this effect here. We know that some of these backs will break out and rush for a whole boatload of yards. But my model isn’t going to take risks on backs to do that so it discounts the yards for the whole lot.

The key to my running back model is really to use each player in context. So while LT’s 994 yards may seem too few, when we see him in context, he has the third-highest rushing total. Now that seems acceptable. The same goes for receptions where Brian Westbrook will have the most, even if the total is 33 below his 2007 total of 90.

Without further ado, the list:

Read Running Back Predictions for the 2008 Fantasy Football Season »

Last week I wrote about the Istobe willingness to share with you some of the predictive models that we use in our fantasy drafts. Well, a week later and the Istobe world headquarters is just now kicking the empty pizza boxes to the side and emerging into full sunlight with one hand visoring our eyes. Indeed, it’s been a computer-heavy week running model after model in an effort to win the right to mock our peers.

My quarterback-selection model (like my running back, wide receiver, and tight end models), which I share with you today, is actually a series of regression models that use 27 different predictors to arrive at an estimation of how NFL quarterbacks will fare this year in the following categories:

• Completions
• Passing Yards
• Interceptions
• Passing Touchdowns
• Fumbles
• Rushing Yards
• Rushing Touchdowns

I used these categories to derive a fantasy value based on my league’s scoring system and urge you to the do the same.

The Highlights

1. Tom Brady will have another big season. He’ll ease back on the touchdowns and increase his interceptions. But the yardage? Off the hook. Check out those 5200 yards he’s going to pass for.

Read Quarterback Predictions for the 2008 Fantasy Football Season »

What’s not to like about fantasy football? It raises the importance of the unimportant football games (i.e., those games in which your team is not participating). It allows for some of the finest trash talking on the planet - with few repercussions for losing the Sunday match-up as long as you win the weekly war of words. And, most importantly, fantasy football gives Istobe yet one more opportunity to crunch data in an effort to predict outcomes. So in the next two weeks, we plan on crunching the last seven years of NFL player data in an attempt to predict this year’s fantasy workhorses.

Read Get Ready for the Ignite Fantasy Football Draft Predictions »

Just finished getting through the backlog of my Don Dodge RSS feed today and I’m happy to report that venture capitalists seems to think that businesses like Istobe are about to break out. Let me qualify that. Venture capitalists seem to think that using data to improve e-commerce is an industry that clearly needs some maturing and that maturing time is nigh. Istobe represents that maturing, combining hundreds of models and data integration routines into a package that lets you target the right customer with the right product at the right time.

Investors believe that the maturation in this industry will occur in the next five years. Well, so does Istobe. We believe that it’s time to put your data to use. If you don’t use it, data is no more than the new shelfware: that software you just had to have before you realized you lacked the in-house talent to unlock its value. Istobe is your outsourced in-house data analysis talent that lets you ask simple questions and get answers without analysis, questions like: I need to sell this overstock of shirts, to whom should I market them? In reply, Istobe gives you a list of your customers that are most likely to buy your shirts and the probability that they will buy. This is the new paradigm in predictive modeling that Gartner calls the data mining packaged application. Let’s just take a quick look at the moment in time at which we are poised.

Data collection methods are clearly refined

Nowadays, everyone is sitting on a pile of customer data that they don’t know what to do with. As Rob Hayes, partner at First Round Capital, says in the Stefanie Olsen article from which Dodge draws his inspiration, “Everyone talks about all the data that’s being created and how valuable it is, but the way you make it available is by doing something actionable with it.” The glut of data is due, in part, to years of CRM implementations and the current CRM zeitgeist. You can’t turn around without being inundated by a flood of marketing for “next generation” CRM systems. In fact, I used to be one of those marketers at a not-too-small company that builds Dynamics CRM.

Of course, online everything has made data more prevalent as well. Returning some kind - any kind - of functionality in exchange for your profile is no longer new hat. Every widget and social network known to man requires you to divulge information before you start using it. And then it collects your clickstream as you use the app. Heck, I’ve got at least 50 different login/pass pairs that I need to remember now.

Purchases, as well, fall into this category. With online shopping increasing at a terrifying clip, all of your purchases are more seamlessly collected and tied to your profile and your clickstream, meaning that now, more than ever; pre-purchase behavior and purchaser characteristics - on a large scale - are at a marketer’s fingertips.

Data analysis tools have matured but haven’t turned the corner

So there are various types of data out there right now that need tying together and, ultimately, analysis. But have the tools to merge and make sense of that data improved? Not appreciably. Really, when you get right down to it, the tools used to perform data analysis are still catch-all tools that can build any model you want or merge any type of data you want. But they can’t help you with specific business problems. In other words, the tools exist for database experts (data integration) and PhD statisticians (statistical modeling tools).

For years, these tools have gotten easier for experts to use but haven’t gotten any easier for business users. This means that your $300K worth of in-house data integrators and PhD statisticians have become slightly more productive over time but translating their language into the language of business is as difficult as ever. And turning the data they spit out into a meaningful business strategy is just as tough.

Recently, I’ve noticed that the vast majority of articles and blogs out there on marketing and customer analytics primarily focus on how to use data-mining to increase revenue and cut costs in B2C oriented companies. Today, I wanted to take 5 minutes to outline some ways B2B companies should be using analytics to increase their top line while minimizing costs.

Knowing Your Leads

Many sales people will tell you that they already know which companies and decision makers they should be selling into. The marketing folks will tell you they already know who matches their ideal lead profile. Funny part is that in many cases these two ideals don’t match up.  A recent marketing research report concluded that in nearly 50% of companies out there, marketing and sales do not have a direct view into what the other side is doing. Data analytics can clearly help bridge this gap by following a customer through - from lead to purchase to maintenance - showing which leads result in the most profitable customers - not just in terms of initial sales, but also in terms of potential lifetime sales. It’s the best way to make sure both departments stay on the same page and the entire company increases its overall hit rate.

Better Cross Sell and Up Sell

It shouldn’t surprise anyone that better targeted email results in better cross-sell and up-sell rates. Personalization is a hot topic in the B2C space, but there is no reason why B2B companies shouldn’t have the ability to tailor their emails to match the company profiles of their customers.  A CTO shouldn’t be receiving the same sales pitch as the marketing folks and someone in the energy business shouldn’t get an email that is designed for online retail. These seem like easy rules to follow, but the trick comes in knowing how to segment these customers by industry, contact, and previous purchase history to minimize the work needed to pitch the right offer at the right time.  No one has time to draft up 100 emails to each type of customer - which is why predictive analytics is critical for optimizing the segmentation process to forecast which customers are going to want similar products in the future.

The True Cost of Support

Many times support personnel get relegated to the back office and we tend to forget about them when factoring in the true costs of sustaining a relationship with a customer. If a customer spends $1,000 on one of my products, but then uses up 100 hrs of one of my support staff, are they still a profitable customer? Factoring in and then forecasting support costs can be an extremely valuable way to judge the lifetime value of a customer and help management decide whether they should renew contracts or continue maintenance agreements.  Working this back into lead generation can also help marketing make sure they focus on customers that will not only buy product, but also won’t cost the company an arm and a leg once that customer comes onboard.

These are really just a few ways B2B customers can make use of data analytics. There are many more, and next time I’ll try and focus on some leading edge cases out there that show how B2B companies are using analytics to increase revenue and cut costs.

I noticed that the RRW Consulting blog alluded to an article on Friday that I have been promoting to my peers: a research report by the Aberdeen Group (abstract here) that discusses the importance of email personalization. The one-to-one marketing emphasis in the article is precisely the kind of email targeting that we espouse here at Istobe. Today, I want to expand on one aspect of the Aberdeen report that we spend extra time on at Istobe: the importance of the buying cycle in determining what kind of email message to send your customers.

In the Aberdeen article, Ian Michiels mentions that web analytics provide great clues to assessing where customers are in the buying cycle. For example, if a customer invests a vast amount of time clicking about a product group, that customer is likely doing research and is in the market to buy a product in that area. A discount offer, Michiels says, would likely get this customer - who is now highly qualified and advanced in the buying cycle - to act on their desire and make a purchase.

I totally agree with this sentiment. But as Chris mentioned in detailing his experience with GPS systems at Amazon, there is another way to do this. Customers can clue you into what they want via their clickstream. But even if you don’t have clickstream data, transaction histories, once supercrunched, can give you a leg up on finding customers who will likely buy next. In other words, this supercrunching can help you locate the customers that will likely buy before they locate you.

How does this work? Well, other customers have come before them and laid out patterns that aren’t perceptible to you and I but are very perceptible to Istobe’s predictive models. Istobe’s models throw out those customers that are not likely to buy again and then work with those who are. From there, Istobe’s models assign the products that are likely to be purchased by these likely buyers.

I won’t argue that this method is more statistically powerful than clickstream data, which is a solid indicator of future behavior. But I will argue that clickstream data takes vast amounts of resources to capture and use, a difficult proposition for online retailers who are just dipping their toes into analytics. And using transactional data to predict who will buy next is a more proactive approach. So what do you get from that proactivity? Probably a two- to three-month head start on your competition. You can focus on targeting your “most likely” customers with act-now offers while your competition waits for these customers to visit their web site.

Something that we constantly talk about around Istobe is why more small companies aren’t interested in implementing customer analytics. Especially given that these companies are pressed now more than ever to keep up with big companies that run customer analytics algorithms as part of their hourly routine. And in an era when a tome like Supercrunchers proclaims a new dawn for predictive models in business. Even with a value proposition that demonstrates a solid return per thousand customers, we often have problems convincing customers that our mojo is good and that data mining is more than just voodoo. Well, today I prepared a really simple sample revenue calculation to quantify the value of Istobe’s predictive models. Take a look at how much more an Istobe-based email campaign makes on a yearly basis.

The rule of any business venture is it must either make money or cut costs for clients. If MIT Sloan didn’t drill this into us, we got a refresher into this the other day when Doug, Chris, and I met with Chris Merrill, founder of Thrive Networks, The Orchid List, and Owner of Ass Industries. He broke that down pretty plainly for us when talking about our need to educate our customers. Well, here is as plain a case as I can make for the revenue increase that Istobe brings to the table when we mine customer data to determine what product they want to buy next.

The Revenue Increase

Assuming a company that emails 500,000 customers weekly, 50 emails a year, and $1.90 per email open, the Istobe predictive models allow a company to make $412,300 more than just basic email blasts and nearly $80,000 more than a baseline targeting approach. There is a full calculation below but let’s talk about what some of the numbers are before you examine them.

The Setup

First, let’s take a very conservative number of 4% as the increase in predictive power of our models over a general baseline. In this case, we assume the baseline to be the company offering the same product that a customer already bought. As creatures of habit, consumers are typically likely to buy the same thing from a company that they just bought. Let’s take dog food at a grocery store for example. Whoever buys dog food probably has a dog and there is a good chance that they’ll keep buying dog food. It’s a safe bet. The same goes for internet retail. If I buy a shirt from an internet retailer, it’s likely I’ll go back there to at least look for another shirt - assuming I didn’t hate the first shirt and the customer service was adequate. Now, the pitfalls of offering the same product is something I won’t go into in detail here; suffice it to say that offering the same thing that a customer is likely to buy again and again from your company is not the best business move. Really, your company should be looking to expand the share of your customers’ wallets. And this means offering them new products that they currently buy elsewhere. Like I said, a topic for another day. So Istobe does about 4% better than offering the dumb alternative, which is offering the same product again. 4% doesn’t sound like a lot but let’s play this through.

Now, let’s assume that a client uses our predictions to send out targeted email. It’s been assumed that normal behavioral targeting enhances your open rate by 40%. This means that just offering your customers the product that they bought last time increases your open rates by up to 40%. Well, I’m not so sure that I trust those studies so let’s just play it conservative and say that targeting spurs a 20% open rate increase. What that means ultimately is that Istobe predictions improve the targeting - the 20% increase - by 4%.

As a base open rate, we’ll use 3.5%. That’s a pretty conventional rate when sampled from our customers. When you apply the Istobe predictive power (4% increase) and the lift from general 1:1 targeting (20%), Istobe’s open rate is 4.4% and the baseline targeting rate is 4.2%. The normal open rate, remember, is 3.5%. This means that, if we take 1000 customers, Istobe will get 44 opens, baseline targeting will get 42, and the traditional email blast will get 35.

Here, we’ll use $1.90 per open as the amount of money that our fictitious company earns per open. Ultimately, not all opens are sales and this figure essentially backs out the complications associated with the website, ordering mechanisms, sizing problems, etc. In other words, it’s a way to understand how much each open is worth on average.

When you take into account that each email open is worth $1.90, and that Istobe has 44 opens per 1000 customers, Istobe’s predictions make our fictitious company $82.99 per 1000 customers per week. Baseline targeting makes $79.80 and regular email blasts make $66.50. Calculated for 500,000 emails per week for the course of a year, we get the figures that I began with at the top.

The Calculations

When we talk with customers about improving the success rate of cross sell and up sell opportunities on their existing products, many reference Amazon.com as the retailer they believe is doing the best job in the industry.  While I agree that Amazon does a fine job with following up on self-generated leads, I don’t agree that they do an especially good job of anticipating or prompting additional purchases.

Let me give you an example. I bought my mother a cookbook three weeks ago for her birthday. I wrote out a gift card, opted for the additional gift wrap, and shipped it to my parents’ home. Additionally, because I’m particularly anal when it comes to data organization, I tagged it within Amazon as a present that was bought for my mother. Now, knowing all of that info, any guesses as to the next product Amazon decided to pitch me on? Cookbooks. I can maybe understand the tendency for someone to purchase the same product again after purchasing it once…maybe…but in this case, really? After doing everything I could to signal to their system that the purchase wasn’t for me (different ship to name, different address, tagged as belonging to a different person), I still get an offer for cookbooks?

Now, don’t get me wrong, Amazon does do many things right. They are one of the best on following up on customer clickstream data. A recent visit to the site to research GPS systems for an upcoming trip resulted in two emails offering deals on GPS systems I actually might be interested in.  My problem with Amazon, and more specifically, the perception of their analytics is that their email marketing model is based on being reactive - the customer is responsible for coming to the site and narrowing down the type of product they want to buy - and Amazon reacts to this behavior with an appropriate offer.  But for those retailers that aren’t Amazon, who can afford to sit and wait for customers to tell you what they want?

The key to predictive analytics is being just that… predictive.  Most retailers don’t have Amazon’s budget and need to have better information about who is going to buy next month or next quarter, and more importantly, what they’re most likely to buy. And for most, the surprising fact is it’s not much of a leap to get there since the majority of companies are already sitting on all the information they need to better target email blasts by matching offers with customers who are willing to buy.

I praise Amazon for their ability to follow up with customers that express interest in certain products, but as far as predictive analytics….there are better solutions out there.