Category Archives: marketing

5 Mistakes you make when you leave the science out of marketing

Posted on by

Marketing is naively thought to be mostly art and very little science. While it is true that there is are elements of inspiration and creativity and passion involved, the balance of an effective strategic marketing role is heavily in favour of science.

As a further point to consider, I put forward the proposition that much of really great science involves inspiration and creativity in passion in more than equal measures to a successful marketing decision. Newton’s development of the laws of motion and gravity, Copernicus’ solar-centered world, Pasteur’s painstaking experiments to support and understand germ theory are all well known examples of brilliance and flair combined with method and rigour.

But where does science contribute to marketing? Is it possible to reap the benefits of logic and analysis and rigour without damaging the creative process?

The answer is “absolutely without a doubt” for numerous reasons. I will touch on just one in the next few paragraphs to demonstrate the idea.

Introducing analytics

The most commonly thought of analytics when it comes to marketing is customer analytics. Better understanding of customer behaviour, preferences and ultimately buying decisions is enormously valuable. Take what was done in the past, compare the success rates of the different initatives, and stop doing the ones that don’t work.

Any organisation can benefit from understanding what works and what doesn’t, and shifting resources to those functions that work. Good organisations also understand the value of play and experimentation, and will continue to allow an element of trial and error. Truly excellent organisations combine experimentation with analytics to truly understand on a measurable level which experiments work and which should be tossed.

A real life example of the place of analytics

Let’s consider a very specific example. An old university friend of mine has started a new venture with a unique offering that clearly means a great deal to him. He has the passion, and presumably the product, to make his idea a success. He also had the good sense to plug into social networking platforms such as Facebook to spread the word of his new website and associated content. So far we have an excellent platform for success.

Mistake #1: Ignoring pre-existing science and analytical results

However, the design of the website appears to have been performed without understanding the hard, measurable evidence from a range of pre-existing studies and material. The website makes it difficult to buy. A long slide-show intro precedes access to the main page, frustrating regular visitors to the page (the intro cannot be skipped) and severaly damaging the ability of his site to be spidered and highly ranked by search engines.

So several mistakes have been made by disregarding the clear evidence that has been accumulated through analysing customer behaviour on similar projects.

Mistakes #2: Not performing analytics on web page at the outset

An excellent first step in understanding how customers will interact with your sales channel is to watch customers interact with your sales channel. Before a site goes live, invite some representatives from your target market (friends and family will do in a pinch if budget is tight, as long as you are confident they will give honest feedback).

Watch them interact with your website (or other sales and information channel). Where are they confused? Do they ask many questions? You won’t be there in person for most of your customers. What do they say is good, what do they say is ugly. If one guinea pig says something doesn’t work, that could be personal preference. If all 3 or 4 give similar feedback, the scientific evidence is mounting and a wise marketer would make changes.

This can be a very quick and easy, but amazingly valuable way to understand the strengths and weaknesses of your approach. Don’t assume you are just like your customers.

Mistakes #4: Not starting to collect analytics and data from the start

It is so easy to collect useful information, if you plan it in from the start. Once the system is set up and the process is working, invaluable information will flow with every visit, every call, every surf and every purchaser.

Not setting up to collect data is usually the first sign that the marketer doesn’t understand the value of understanding the customer.

Mistake #5: Thinking science cheapens the experience

Perhaps this should be mistake #1. Many people with great ideas feel that their ideas should sell on their own merit. They view logical, analytical understanding of customers to be beneath them. If the product is good, if customers will benefit from purchasing the good or service from you, then you owe it to them to make it as easy as possible for as many as possible of them to effortlessly find their way from oblivious potential customer to satisfied repeat customer.

If your aim is to build the perfect mousetrap, perhaps it is worth finding out what customers want in a mousetrap, where they like to buy it and how they like to buy it.

Will an individual investment product development role help towards actuarial consulting career?

Posted on by

An old contact mailed me a couple of days ago asking my view on two possible actuarial positions. He was interested in whether either would be suitable preparation for a consulting or actuarial consulting role in future. I thought the answer might be of broader interest, so I’m copying it, with a little editing to protect the innocent into a couple of blog posts.

Please note this shouldn’t be taken as categorical always applicable advice – just some thoughts:

Insurers are facing stiff competition in the investment space, both on the individual and corporate side. Competing against investment banks and asset managers, often on more direct terms than in the past, and with the albatross of poor reputation, sluggish reaction to changing market conditions and often a higher cost base mean that insurers need plenty of help in this area. Depending on the specific role, you can get involved in anything from market research and customer needs analysis, creative product design, model coding, profit testing, distribution analysis, customer segmentation and analytics, financial reporting (IFRS, FSV/PGN104, EV, EEV, MCEV, Solvency II etc.), administration system design and implementation, liaison with distribution force, sales team training, product profitability monitoring, policy documentation wording etc.

There is plenty of scope to gain breadth of knowledge, but it is likely that you won’t be doing all of this. I’d check out what they envisage you actually doing, see what is in your “performance contract” if this exists and see whether it is a narrow or broad role, and whether it can become broader over time.

Aside from its relevance to an actuarial consulting career, this is an interesting space to be involved in at the moment. Around the world, insurers are only touching on possibilities of product design and customer information and analysis. South African insurers are probably a little further behind the curve.

Good luck!

Why premium size matters (more than you think)

Posted on by

Most people involved with insurance recognise that more premium is better (ceteris paribus of course). This is usually true (and occasionally not) but while some of the reasons are obvious, there are a variety of more subtle factors to take into account. This post will cover many of these factors, and point out a few cautionary tales around seeking large average premium size above all else.

When is value created?

For a particular product-type, it is usual for larger premiums to be more profitable than smaller premiums. By profitability here I mean the increase in shareholder wealth resulting from having sold that additional policy. The value creation at time of sale arises from:

  1. A customer relationship has been confirmed and cemented through an agreement to do business for a few months or many years. The customer relationship was already in the process of being developed in the period up to the sale (from broad advertising campaigns, brand-development, specific distribution channel contact and the quotation process). However, this is also true for any other industry, so we will restrict the analysis in this post to the “point of sale”.
  2. This customer relationship means that for short-term or annually renewable business, there is a non-zero probability of renewal, and this probability is likely to be higher than the probability of a random individual with no previous contact with the insurance company buying a new product under the same conditions.
  3. The costs of renewing an existing policy are usually lower than those of creating a new policy. (Policyholder and risk details are already captured on the system, the sales process is quicker, legal and regulatory compliance (for example, around identifying customers) is already complete and payment details / credit checks have been performed.
  4. For life insurance business, a long-term, legal contract has been entered into. Traditionally, these contracts can be cancelled at the option of the policyholder (usually with a fair and sometimes controversially unfair penalty). In spite of the cancellation option, signing a long-term contract provides some evidence that the policyholder has an intention to enter into a long-term agreement with the insurance company.

This list isn’t exhaustive, but it covers some important bases.

So why are larger premiums better?

Larger premiums are more profitable because:

  • Some marginal costs are fixed per policy
  • Larger policies are usually more persistent
  • Larger policies usually imply greater wealth, which usually means lower morality (check below for caveats!)
Some marginal costs are fixed per policy

Many actual marginal costs really are fixed per policy:

  • Policy form, posting, printing, filing etc.
  • Ongoing reporting and communication with the policyholder
  • Bank charges related to processing premium receipts and claim payments
  • Calls to the call centre
  • Valuation modelling costs (PC / Mainframe running time, purchase costs, electricity, coding, debugging)

Since the costs are fixed per policy, the greater absolute charges are matched against lower costs yielding a higher margin.

Larger policies are usually more persistent

No question this is subjective, but one only needs to consider the 25% – 50% first year lapse rates on low-income products with small sums assured and small premiums. Large policies are likely to be sold to educated consumers who are less likely to be hoodwinked by smooth-talking commission-driven salespersons.

One can understand logically how this could be true, and the data supports these conclusions as well.

Larger policies usually imply greater wealth, which usually means lower mortality

Fairly standard actuarial knowledge this. Higher income means better access to healthcare for current ailments. More importantly, high income now is strongly correlated with high income in the previous years, which implies consistently better access to good healthcare and thus better overall life expectancy. Moreover, higher income is correlated with higher education. Education is correlated with family having money, which is correlated with good healthcare since birth, which is positive for life expectancy. Certain diseases (particularly heart disease) are related to stress and high cholesterol, which are positively correlated with wealth and income and act in the opposite direction.

Lower mortality both means lower claims experience (for non-annuity risk products) but also means, very marginally, that persistency will be higher since dead policyholders don’t pay premiums. Since a portion of all premiums is earmarked for the repayment of initial expenses, the more premiums paid the higher the overall margin will be.

And what about the impact of discounted rates?

Absolutely right. Higher premiums often attract discount rates, including lower asset management fees, higher allocation rates and lower mortality charges. These cost elements shouldn’t be ignored in the analysis, but experience usually shows that the benefits outweigh these costs. Results may vary!

An element that is often forgotten is medical underwriting. Most underwriting manuals have limits below which certain components of the comprehensive underwriting process are omitted because they aren’t cost effective. Thus, for the largest policies, the underwriting cost are often the highest. Analysis of actual experience and the costs involved for this should provide reasonable estimates of this cost.

One final, even more subtle impact is that of statistical variation. Individual policyholders will die (and we are continuing with the focus on non-annuity risk products here) with a certain probability at each age. Thus, the overall distribution of the number of deaths in a year should follow a binomial distribution, ignoring catastrophes and the slight theoretical correlation between deaths of spouses. Since the number of deaths follows a binomial distribution, we can determine likely variation from the expected number of deaths using basic statistical methods. What this also shows us is that as the number of policyholders increases, so the percentage variation from the mean decreases due to the diversification benefit. I’m not going to go into the detail of this for now – those familiar with insurance should be comfortable so far.

So, ideally we want lots of policies. If we also want to hold the premium constant between two comparable companies (S and B) but where S has small premiums per policy and B has big premiums per policy, then S will have a greater number of policyholders than B and will experience less volatility in financial results through better diversification of risks. You can also think of this like every Rand (or Pound or Dollar or Euro) of benefit for a particular policyholder is perfectly correlated with every other unit of currency for that same policyholder. Either the policyholder lives and all units of currency don’t get paid, or the policyholder dies and every single unit of currency is paid out as the total Sum Assured. Thus, larger premiums make larger benefits make more correlation and less diversification. This slightly unusual way of looking at the problem is what most people are familiar with as concentration risk, except here we are considering concentration within individual policyholders. This increase in risk increases the economic capital required (and often the regulatory capital too) which will likely have a cost to be considered.

So large premiums matter

Most people involved in life insurance will intuitively feel that larger premiums are “better” or more profitable – here are some of the reasons why. Most of these reasons are familiar to actuaries, and if you give an actuary a little bit of time he or she will likely come up with these and some others as well. However, this article has focussed on premium size in the absence of other factors and incentives. I’ll post soon on an example of how the external environment can distort this natural operational conclusions.

Measures, targets and Alchemy

Posted on by

When a measure becomes a target, it ceases to be a good measure.

I first heard this quote when dealing with performance measurement and remuneration structures for senior management. In that scenario, the danger is that you get exactly what you measure rather than the good behaviours related to or driven by the metrics chosen. The measure starts as Earnings Per Share (EPS) growth, which is generally a good thing. However, once management do the maths, they realise that reducing dividends to zero will boost EPS growth, even if it means pursing projects with a return lower than shareholders’ cost of capital. Measure becomes a target; measure ceases to be useful.
More on that some other time – it is an interesting point itself.

Now on to the magic and mystery, and science and great skill, and analysis and mathematics and theories, and occasional quack and snake-oil salesman – Search Engine Optimisation. Isn’t there an argument to say that, if the aim of search engines with their great yet imperfect algorithms is to reward fresh, relevant and useful content for relevant search terms, then the best long-term strategy would be to continue to write and publish fresh, relevant and useful content? No quick wins, and with less of the alchemy involved SEO companies wouldn’t get as many customers, but why isn’t this the best advice for long term traffic and search engine ranking? Rather than pursuing loopholes and quirks in any particular (temporary) search system, the measure should match something more fundamental – being a useful website. Difficult for that approach to need to be changed when Google uses “nofollow” links or omits duplicate stories or starts recognising your “invisible white on white text”.
Ok, before the backlash begins, there are practical lessons that can help search engines. “Obvious” things like “search engines are not people and therefore will struggle to read text if it is really a picture embedded in a fancy Flash animation”. This is probably a bad example – I expect fresh, useful and relevant content appears less often in glitzy Flash clips.

So isn’t it time to take the difficult medicine, and build a brand and loyalty and readership and customers and repeat business and structure value and goodwill by actually earning it?

Markets, unintended consequences and the spam in the Feudal System

Posted on by

Ok, so the title will only make sense within the context of Blizzard’s hugely popular (and financially successful) World of Warcraft. WoW is, very simply, a multiplayer online game (and by multiplayer we’re not talking of 4 players here, but rather millions of players around the world) where players interact with other players and the virtual world of the game.

A key component of the game is that special items can be purchased. The currency is “gold” and gold can be earned in a variety of ways. The least interesting of which is by performing basic tasks and completing basic quests. These are, in general, not very difficult or challenging, but do still take a fair amount of time. The other piece of the puzzle that is relevant to this post is that gold can be transferred from one player to another.

Without too much difficulty, it should be clear that players with lots of time and little money are incentivised to spend their time earning “in game currency” to sell to time-poor and money-rich players for real world cash (i.e. as in US Dollars). The sale of in-game currency in the real world is a free market, so free-market economics forces act on the allocation of resources (time and money) in a way as to more optimally allocate resources.

Picture if you will the far-eastern sweat-shops manufacturing shoes. Now replace the glue and sewing machines with computers, and replace the shoes with in-game WoW currency. Cheap labour comes to the fore and a business is born. In-game currency is the product, salaries (and a bit of computer, WoW and internet expenses) are the Cost of Sales, and real hard cash from time-poor First World teenagers and adult is the revenue.

Two problems:

  1. Is purchasing vast quantities of in-game gold with real-world currency (presumably earned from applying one’s particular real-world skills and talents in gainful employment) cheating? Is there a moral or ethical angle here? What is it?
  2. How do the “entrepreneurs” boost sales? They advertise! How? By spamming in-game players with messages.

The first point is interesting, and worthy of a blog (and quite likely a UN commission as well). The second point that has seen some recent action from Blizzard, who are suing one of the companies behind the in-game spamming. Will be interesting to see how that develops. Slashdot also picked up the story.
But, understanding how the problem arose is clear with the benefit of hindsight. However, I am quite certain that with some basic analysis of the economic forces in the game, and an understanding of consequences, these problems should have been anticipated by Blizzard. Seems like they have fallen one step behind the spammers, which could also be interpreted as the power of the free market.

Oh, and the Feudal System? The company being sued is “peons4hire”.

Pricing and promoting businesses

Posted on by

David Maister’s blog has an interesting post about the pricing and promotion strategy for a small home-based business. In this case, it’s a pre-school looking for a pricing and promotion strategy. As always, David Maister has some useful suggestions and no doubt his wealth of readers will add some real gold in terms of suggestions soon.

The case involves the pricing and promotion of a pre-school in an area where there are “informal” preschools with little or no qualifications, and formal preschools with highly qualified teachers. The different schools have different prices and demand levels.

I approached the problem by applying a model. In this case it was the “4 P’s of Marketing” model. It is by now an old model, but grew out of a very different time when it was assumed that if you have the best product, customers will beat a path to your door. This was followed by the idea of price competition and price being the all powerful tool to sway consumers in their decision to open their wallets.

Most people now realise that these ideas are too primitive. The 4 Ps are:

  1. Product
  2. Price
  3. Place
  4. Promotion

Some people add a 5th P called “packaging” to the list, which may or may not be useful, and could probably be added under product, with the idea that the product purchased is the full basket of utility attached to buying the item, or added under promotion as part of the more traditional selling process. It’s just a model after all.

I like to apply models to problems as it provides a framework to generate ideas. No doubt there are good models and bad models, and models will sometimes restrict thinking and options as well as freeing up the mind to think of new options. (I am also a big fan of brainstorming and other structured lateral thinking methods to generate creative ideas.) But without a model, it can be difficult to get everybody speaking the same language, and it is easy to look at a problem from only a single dimension. In most cases, applying more than one model is even better. Too many models and you don’t spend enough time applying your mind to the thought patterns under each model.

I mention “models” quite a bit on this blog. Here I am describing mental models for approaching problems, rather than computer-based models for figuring out the numerical answer to a problem. Different application, but both useful.

Here’s my response to the pre-school post:

The question was posed about both pricing and promoting, but then it seems that the rest of the comments all related to price. The 4Ps recipe is an old one, but is still a worthwhile place to start for an initial structure.

Product, Price, Place and Promotion

1 Product

A good product is important, but it isn’t going to sell itself. The analysis presents some information on pricing and supply / demand levels for various “products”, but it is unclear to me whether this analysis is systematic and representative or not. As tough as it is, this is something you need to look at through dispassionate eyes at times. Do other parents want qualified teachers? Or do they prefer a less formal approach. Maybe (and without children myself I am purely putting this out as a possible point of view for discussion) parents feel that they are not handing their children over to an institution if the teachers are less qualified and the setup less formal.

2 Price

Price has been given more airtime than the other components so far. Coming from outside the States, I have no feel for absolute levels of pricing. However, maybe this is an area to brainstorm a large array of pricing possibilities (and get David’s readers to suggest many more!). Maybe you offer reduced rates for the first 5 kids to sign-up, or a discount for upfront payment for a 3 or 6 month period. This could help finance the cashflows early on, encourage some interest and early adoption, but also make it a seamless process to go from discounted prices to a premium-priced service. I do agree with David’s point that it sounds like a tough ask to get a full house in a short space of time, but I suppose a preschool with 2 children isn’t much of a fun place for the children to play and learn.

3 Place

Is the place to which you’re moving right for a preschool? How much has the desire to start a preschool affected the choice of area? Are there other areas that meet all the other criteria (the writer’s own workplace, nearby schools, other amenities and “feel”) but are better suited to start a preschool? What sort of catchment area do you envisage? If the children’s homes are spread around, maybe some form of transport service could transform “place” into something more workable. This again needs a firm understanding of real demands of prospective customers (both the children and their parents).

4 Promotion

How you go about promoting the preschool will be crucial. Flyers placed on cars may be successful (I’m not convinced) but face to face visits to families in the area might add a personal touch, a relationship and trust-building touch that will go a long way to settling anxious parents’ minds and showing that you are serious about a quality, professional product based on whatever mix you are after. If your wife believes in a mostly play-based preschool, then that is what you must show. If it is going to be desks and chalk boards and 18 hours per day of advanced mathmetics lessons, then that is what you convey. Personal selling will allow time for the parents to ask questions and get to know the people who will be looking after their children. If the service fits, I expect the price will be less of a sticking point.

If you use a personal promotion strategy, it also gives you the opportunity to receive instant feedback early on, which you can use to adjust your promotion, pricing and even product strategy. Started early enough, it will advise you whether the “place” you have chosen is going to work to. This goes back to my allusion earlier on that you may need some more hard data and careful analysis before you set everything in motion. The odd discussion with friends over at a dinner party (I don’t mean to suggest you haven’t done more serious research than that, but hopefully you understand my point) does not replace carefully considered homework to make the launch a success.

Because the number of successful promotional visits is quite low (20 or so I gather) it is more sensible than if you needed to fill 5,000 seats for a convention!

A 5th “P”?

There is an occasionally added 5th “p” – Packaging. Unless you’re planning to wrap the kids in bubble-wrap to keep them safe, I think we can safely skip this one!

A bit of the science behind Google and PageRank

Posted on by

I’ve said before that I’m not an expert on Search Engine Optimisation, but it certainly is an interesting area where businesses of all sizes can reap definite benefits from applying a little analysis, maths and science to the problem.

Have a look at this article explaining, in reasonably simple terms, some of the background behind PageRank.

Additional Analysis of SEOmoz web popularity data

Posted on by

SEOmoz.org provide some great resources on search engine optimisation (“SEO”). Recently, they performed a really interesting analysis comparing actual site traffic for 25 sites that volunteered their data against indicators from a range of competitive intelligence metrics from sources such as Google PageRank, Technorati Rank, Alexa Rank and SEOmoz.org’s very own Page Strength Tool. The stated goals of the project is described in this quote from their page:

This project’s primary objective is to determine the relative levels of accuracy for external metrics (from sites like Technorati, Alexa, Compete, etc.) in comparison to actual visitor traffic data provided by analytics programs. 25 unique sites, all in the search & website marketing niche, generously contributed data to this project. Through the statistics provided, we can also get a closer look at how the blog ecosphere in the search marketing space receives and sends traffic

You can find the commentary on their updated analysis and also the original article (updated too, I understand).

Now, I’m not yet an expert on SEO, but I do know a few things about data analysis. Whereas their results indicate that none of the measures are particularly useful, I have three points to add:

1 Significance of correlation coefficients

A correlation coefficient does not need to be 0.9 or 0.95 to be significant as mentioned:

Technorati links is actually an almost usable option at this point, though any scientific analysis would tell you that correlations below 90-95% shouldn’t be used.

Roughly speaking, correlation coefficients greater than about 0.7 or 70% explain approximately half the variability in the observed variable (actual page visits). Whether or not this is “significant” depends on the amount of data used to measure the correlation. There are some very specific tests for measures of significance for correlation coefficients – I have summarised the results of one of the standard tests here:

SEOmoz data Correlation Significance Table

Beyond the technical statistical tests though, I would imagine that there is a great deal of value in estimating a large part of the practical popularity of a website (and presumably page visits is a sensible measure of this) through freely available “competitive intelligence metrics”. On the other hand, if you are looking for a near-exact replica of actual visits, then a much higher correlation coefficient is required.

2 Extending analysis to multiple regression rather than single correlations

OK, this does take the analysis beyond the original stated goal, but it is interesting to see how good a model of actual site popularity we can develop based on freely available “competitive intelligence metrics”. But first, it is useful to consider the correlation matrix between all variables (the “dependent variable” and all independent variables). In an ideal regression model, the independent variables will be uncorrelated with each other. On the other hand, if these metrics are any good, we would expect them to be strongly correlated with each other.
SEOmoz data Correlation Matrix
As can be seen from the table above, there are several strong correlations between the independent variables. This can lead to problems with “multicollinearity” for multiple regression technqiues, but since I am trying to keep this post non-technical, I’ll leave that alone for now. It is also interesting that while all the large (loosely defined here as greater than 70% or less than -70%) correlations are positive, there are many negative correlations as well. Thus, some measures appear to be using different information or approaches to provide the metrics. Most interesting to me is that TR Rank and TR Link have a correlation coefficient of -50%. This will be a hint to our multiple regression results…
I decided to use only very basic tools for the analysis so interested readers can perform the same analysis on their own with only MS Excel (generally a fairly weak statistics platform even with the Data Analysis add-in activated). My aim was to find a model that explained more of the Average Visits than Technorati Links by combining several variables together. I had to exclyde Compete Rank and Ranking Rank due to the limitations of Excel’s regression tools. I would measure “good” models by having a high adjusted R-squared, and significant and sensible estimates for individual variables as well. The results of a “good” model (although not necessarily the best since I did fairly quick and dirty model selection) are given below:

SEOmoz data Multiple Regression Results

SEOmoz data Multiple Regression Results Summary

The model has a “Multiple R” (which is intuitively analogous to the normal Pearson correlation coefficient) of 89%, and the model explains 80% of the variability in Average Visits. Other measures of goodness of fit include a high Adjusted R-squared (relative to other models fitted) of 71%, a F-statistic for overall model significance of 9.5 which gives a significance level or p-value of 0.00008 and low p-values for most independent variables included in the model. The intercept itself is not signfiicant, but we leave it in to improve the overall fit of the model. Similarly, while the significance level for Alexa Page Views is relatively high at 17%, it does add to the overall model in terms of fitting the data well.

SEOmoz data Multiple Regression fitted model

Again, very interestingly but not surprising by now, many of the coefficients are negative. This implies that, at least once adjusting for the other variables, these measures are associated with lower rather than higher Average Visits. This suggests more analysis and more data is needed to understand the dynamics here properly!

3 Quality and quantity of data
This leads me to my final comment. 25 Websites, while great to have even this much data, is not really anywhere close enough data to analyse this problem. This isn’t because of the small size of 25 sites in relation to the total available websites on the ‘net, but rather to do with the spread of sites across the different types of websites and the potential to fit the model too closely to the exact data provided rather than to some underlying reality. Again, this is a difficult area to discuss correctly and thoroughly without becoming very technical so I’ll leave that well alone too.

Final comments

This analysis and presentation of results is very lite for something this interesting. There is an enormous amount more that could be done with time, energy, more data, and, for my part, a better understanding of how each of these competitive intelligence metrics are intended to work. I’d welcome any comments on what analysis would be desired (time-series? Non-linear models? More detailed regression? Rank correlation?) and whether there is any chance of getting more data. I’d be very happy to dig deeper and post the results here and/or directly on SEOmoz.org