The other reason for Google+

I'll start this one with admission; I like and use a lot of Google's products. I've got an Android phone, have been singing the praises of Google+, Google's my default search engine and GMail is fantastic.

Like US antitrust regulators though, I'm starting to wonder if Google might have too much power. If Microsoft had a case to answer in the way that Explorer was been bundled with Windows, then wouldn't Google have similar issues with the increasing integration between its products?

Google has a large suite of products, despite recently closing down Labs and many of them are tied very closely to its search engine.

Search Google for any term that could reasonably return a map and you'll get a map included in the results. A Google map, naturally.

That's fair enough; I was looking for Leeds and Google fetched me a map of Leeds. Maps can probably be included in a legitimate list of the things I might have wanted. Unless you really want to get picky (and if you're Streetmap, then you probably do,) all Google's really doing is returning a graphical result rather than a text based one.

The trouble with this type of justification, is that you can push it into almost any sphere that the web touches. And the web now touches virtually every part of our lives. If I want to do anything then you can say that I'm looking to do it. Which means I'm searching for it. Which means it's a legitimate product for Google to develop and cross promote from its search engine.

This is exactly the argument that Eric Schmidt is pushing with regulators in the US.

"[W]hat is crucial to understand is that universal search results are not separate 'products and services' from Google.

Rather, the incorporation of thematic and conventional results in universal search reflects Google’s effort to connect users to the information that is most responsive to their queries.

Because of this, the question of whether we 'favor' our 'products and services' is based on an inaccurate premise.

These universal search results are our search service — they are not some separate 'Google content' that can be 'favored'."

(Eric Schmidt. Quote borrowed from The Register)

Google's search market share in the UK is over 90% according to Hitwise. That's a hell of a lot of potential abuse of a dominant market position. I'm not saying that Google is abusing its position - the legal work on my house move is costing quite enough - but if everything Google builds can be integrated into search because it's all one product, then where do you stop? Taking a broad definition of the term, virtually everything starts with a search.

If want to know about a location then I need a map, Google has maps so Google directs me to its own maps.

If I need a flight, Google has a new flights product, so I can be sent there rather than to Skyscanner.

If I'd like to call someone, Google has phones and voice and video chat.

It's difficult to think of any service-based category that Google couldn't decide to enter, develop its own product, cross-promote it from search and use that same Eric Schmidt argument as a justification. Google Legal? Google Estate Agents? Music? News? Why not? You're searching for information and content.

As much as I like Google+, I missed one of its primary benefits to Google the first time around and it only became clear when the black product bar arrived, that now sits on top of just about all Google products.

By closely integrating their product offer - essentially by making it all one product - Google are playing the same game that Microsoft tried to play with Internet Explorer. Compare Schmidt's argument above, with this Microsoft justification for bundling Internet Explorer with Windows.

"Microsoft stated that the merging of Microsoft Windows and Internet Explorer was the result of innovation and competition, that the two were now the same product and were inextricably linked together and that consumers were now getting all the benefits of IE for free"

(Wikipedia link)

Sound familiar?

Microsoft ended up in a compromise with regulators, that was likely a much better outcome for them than if they'd just stubbornly refused to un-bundle Explorer from Windows.

Google seem to be playing the same strategy: Integrate your products so closely that you can argue they're not actually separate products at all. In that context, Google+ needs to be a window onto everything that Google does. It also explains why you'd ruthlessly kill off Labs, which might otherwise be cited as hosting examples of off-shoot products that have nothing to do with search.

Google won't get away from an antitrust investigation completely unscathed but it seems to be a good strategy to ape Microsoft and try to avoid a ruling that's heavily weighted against what they want to do as a business.

For their part, US regulators need to recognise Google+ for what it is: not just an aggressive move into social, but a very clever defensive move to counter a future antitrust ruling.

Book Review: Moneyball

Continuing Wallpapering Fog's series of occasional book reviews, we have Moneyball, a book about baseball statistics and a team called the Oakland A's that I picked up on the recommendation of @AdContrarian.

I know very little about baseball, other than that I watched a game on TV in a hotel room in New York once and found it quite dull. Before reading Moneyball, I'd never heard of the Oakland A's, but none of that matters. Moneyball is a stunning piece of work.

Baseball is the background for a story about how to change a business. The huge number of games that get played and the set-play nature of the games make baseball a statistical goldmine, where a few amateur analysts had noticed that a lot of long-standing, established wisdom about the game, was wrong.

One team - The Oakland A's - take this knowledge, which was freely available to anybody with an interest and set about building a team based on what they can prove about the tactics and the types of players that win games. They turn on its head the idea that the team with the highest paid players will always come out on top.

Michael Lewis works potentially dry statistics into a fabulous narrative, interspersed with the life stories of Oakland's oddball players, who don't look like athletes and would be rejected on any traditional evaluation of whether they're suited to the game. Overweight, old, injured and with bizarre throwing actions, they're mainstream baseball's rejects, but they've got stats that say they can hit...

As I read Moneyball, comparable problems from business, marketing and other sports kept jumping out. You find that you break from the page to wonder if the management at Stoke City FC have read it, or to curse (having worked for a year at EMI) that more people in the music industry haven't. You can't help wondering how many of our own established marketing practices are wrong and which ones we could prove definitely are. Baseball's brimming with statistics and yet the task of breaking established ways of doing things is incredibly hard, even when the evidence is staring you in the face. Marketing's a black art at the best of times, where it's much harder to produce the battering ram of hard stats that at least point the right way.

For me, as a statistician, Moneyball inspires, by showing just what can be achieved by dispassionate analysis and is daunting in its illustration of just how hard you have to work, to make the changes you've proved need to be made. Baseball went thirty years before anybody with money and control of a team paid attention to the hard evidence that established tactics and the usual metrics that were used to value players actually harmed your chances of winning.

Once one team picked up this knowledge and started to apply it (via a General Manager who couldn't care about who he needed to fire, intimidate or cajole to get his way), years went by before other managers started to ask how they were being consistently out-performed by a team with only a third of their player budget.

In short, read Moneyball. You don't need to know anything about baseball (though a little understanding of a few key terms, like base stealing, helps) and I promise by the time you've finished it, you'll want to make changes to the way that you work. It's the best non-fiction book I've read this year, by miles.

How to increase your marketing blog traffic

Want to know the secret formula for writing successful marketing blog posts? Then you've come to the right place.

If nothing else, this graphic might explain Wallpapering Fog's frequently disappointing traffic figures...

Chase success

You see this deceptively simple marketing question rear its head a lot, both in agencies and client side. I'd love to know what answer is taught on marketing courses, or even if a straight-forward answer is taught at all, because marketers seem to jump to the wrong conclusion so frequently.

The basis for my answer comes through in statistical models of marketing response, but also fits with a simple view of what can be achieved by marketing. For me, these are all fundamentally the same question:

"We've got a loyalty card and our card-holders don't often redeem their points for certain of the rewards on offer. Should we feature those rewards more prominently in the brochure to boost their take-up?"

"Most of the shoppers in this store come from the east side of town. Should we run some marketing in the west, to let those people know we're here?"

"Some of the holiday packages that we offer are selling much faster than others, should we 'fix' the under-performing ones by marketing them?"

The question being asked in all three of these is, "should I try to persuade people to do something that they don't seem to want to do, by marketing to them?"

The answer - nine times out of ten - is, "no; absolutely not."

It sounds obvious when the question is phrased like that, but think about how many times you've been asked to market an under-performing product, in order to 'fix' it.

Whether a product is selling or not is the single, best piece of market research that a marketer can lay his hands on. If it's not selling to a group of people, then they don't like it. Telling people about products that you already know they don't like, is just throwing your money away.

The basis for trying to boost an under-performing product comes from an over-estimate of the magic that marketing fairy dust can sprinkle onto a product and also a belief that the product you're selling should fundamentally appeal to a wide audience. Both of those come from being too close to the product that you're selling. People who aren't close to your product, who really don't care about your product but are being asked to buy it, have already told you what they think.

The only exception to this rule is if you can genuinely say that people might not be buying your product because they don't know about it. Before you decide that lack of awareness must be the issue and splash the cash on some marketing though, look at the products you've got that are selling. Do people really know that much more about those?

The reward card example I used above is a real one; the marketing team thought that more people should want to redeem their loyalty points for music CDs and wanted to push that message. Unfortunately, up to that point, music CDs hadn't been starved of marketing coverage compared to the other options available and people still weren't choosing music in large numbers. That's a loud and clear piece of the best market research money can buy. You offered free CDs to millions of people and only a few said "yes, please."

It's the same for the store that appeals to only one side of town. People on the other side either can't get to your store easily, or they don't want to. You can tell them all about where you are and they'll just keep on ignoring you.

For me, when a product is under-performing, you've got two options.

1. Save your money and stop pushing it.

2. Find a way to change the appeal of the product. (That's likely to be a product solution before it's a marketing one)

The reverse of all this gives a key piece of marketing advice. If you haven't completely saturated the market, then always chase success. Support your biggest sellers and make them even larger. People are telling you that they like these products - not just by saying nice things in focus groups, but by spending their own money on them. Your marketing will be much more effective if you listen.

Never let anyone tell you that it's too complicated to explain

Sayeeda Warsi was on BBC Question Time last night, giving the Conservative viewpoint on quantitative easing.

Baroness Warsi has a remarkable ability to get my back up at the best of times, but after an explanation of QE that didn't really make sense - to this economist at least - where she said that it "feeds demand... will keep interest low, mortgages low... and gets this country moving again," she went on to conclude that "it's complicated".

A lot of things are complicated. Dark Matter is complicated, but I saw a perfectly watchable Horizon programme about it last night. Econometrics is complicated too, but I don't try to sell it to clients, by rambling something about statistics and then saying "it's complicated".

Shortly after our ability to understand quantitative easing was dismissed with "it's complicated", @Natt tweeted this little gem.

"It's not complicated. You're just devaluing everyone else's money to keeps banks afloat."

And in one short sentence, swept away a tidal wave of bullshit.

Never, ever let anybody tell you that something is too complicated to explain. They're either avoiding telling you the truth, or as Einstein famously said,

“If you can't explain it to a six year old, you don't understand it yourself.”

Do you ever see a TV ad...

... and suspect that a creative had an idea that they've been just dying to dying to use somewhere? Anywhere? Pitched all over the place before finally, a brand agrees that a cartoon lady adrift on the ocean, who finds her cartoon twin on a desert island just screams "you really want to buy our tea."

It is a nice cartoon. Maybe I'm just too cynical.

Dashboard software - why we chose what we chose

Some time ago, I wrote a post that laid out what I want from a piece of dashboarding and data visualisation software. We had just finished going through the process of selecting a software platform and I'd intended to lay out over a series of posts the different options that I looked at and explain how we made our final choice.

Apart from a review of Excel Services (summary: it's sort of ok, but you can do much better for the money), the series didn't get very far. Due to the overwhelming volume of emails Despite a complete lack of emails accusing me of being a lazy so and so who doesn't finish what he starts, I'm going to quickly wrap up the conclusions here.

First, the contenders. Some of these I looked at for longer than others and I can guarantee a few first round knockouts.

Ladies and gentlemen, in the blue corner etc.

Tableau

Spotfire

Qlik View

Excel Services

LogiXML

Microsoft Analysis Services

Pentaho

Microstrategy


We'll start with where I started the process - looking for a cheap but capable solution, that would work for a medium sized business. I'm not looking at six figure software budgets here, or at running the back end for the Tesco Clubcard.

With that in mind, I went looking for an open source solution and found Pentaho. There are other 'free' platforms, including customized Google Docs, but all require programming and we need a drag-and-drop front end that analysts can use to build dashboards. For that requirement, Pentaho seems to be what's available.

First impression is that it's not bad at all. A little rough around the edges as you might expect from an open source project, but I had some simple dashboards up and running on a demo server pretty quickly.

These demos were enough to show a few senior managers and to explain why we should be buying a dashboard platform rather than building our own, but then things started to go wrong. If you're going to do anything more than a few simple charts, then you're going to end up doing some programming, which means as an analyst that you need IT's help. IT's help brings change requests and (long) development time-lines and all the joys of not being able to put data labels on a bar chart by yourself. Scheduled refreshing also looked like it was going to be trickier than it needed to be.

Once you're building and customising with your own IT department, you're incurring extra costs, so it was off to the commercial software world to see if off-the-shelf was a better idea. It was.

When you head for a commercial piece of dashboard software, you find out that there are at least tens available and probably hundreds. Most are rubbish. You also get into some very frustrating conversations with salesmen about licencing costs, but more on that later.

Let's eliminate a few contenders quickly here, pretty much as quickly as I dropped them during the assessment process.

I've already reviewed Excel Services in detail so won't go through that again. For the price, it isn't good enough and it's awkward to use.

Analysis Services is much more capable, but it's got that Microsoft feel. You know what I mean; it makes things hard that should be easy. Options are hidden, it's all tied into SharePoint (which I hate) and you end up writing loads of custom SQL - which will be a nightmare to maintain - in order to work around things that Microsoft either doesn't want you to do or hasn't included as standard. The final dashboards also feel like a tool for analysts rather than a tool for marketing managers and if you're not at least a little geeky, they can be quite intimidating. Rejected.

LogiXML and Microstrategy were very much the also-rans in this race. LogiXML, because it's bloody awful and Microstrategy because it's just not good enough to be up with the front runners.

Here's a LogiXML demo dashboard. Apparently they're proud of it and if you can't see why that's a problem then go ahead and buy a copy! I'm being deliberately harsh because they keep sending me junk email. Petty revenge is sweet.

Microstrategy sounds pretty exciting when you read the blurb on their website. The demos aren't front and centre on the site though and you have to dig a bit to find them. That's a warning sign for me. If a software company knows that it has a good product, you'll find links to the product itself featured prominently. Microstrategy has a lot of claims about what it can do on the homepage and then a few clicks and some searching around takes you here...

I'm sure it works ok, but the looks are five out of ten at best. I did download the demo but it took ages to install and the dashboards are slow to load on a browser too. All in all it just didn't feel like a smooth, slick piece of software and I couldn't see the visuals impressing a marketer. Another one bites the dust.

Now we're onto the serious contenders: Spotfire, Tableau and Qlikview.

Spotfire first, because it didn't make it quite as far as the other two. Spotfire is undoubtedly a tremendously capable piece of software but at Brilliant Media, we're building dashboards for marketers and they need to present data, more than they need to make it available for investigation by an analyst.

With that purpose in mind, Spotfire suffers from some of the same issues as Microsoft Analysis services. It's less awkward than Analysis Services, but in use it feels like it was built for an analyst. It's the sort of software that would be great to sit on top of a loyalty card database for an analytical team to use as their tool for day-to-day interrogation. It's not what we need to present relatively top-line data to marketers.

Take a look at the demos and you'll see what I mean. It's good software, but not for us.

Spotfire has another major problem and that's Qlikview. If I wanted what Spotfire does, then I probably wouldn't buy Spotfire, I'd buy Qlikview. It's an amazing piece of software.

I've been playing with Qlikview demos on and off for a few years and it has a few stand-out characteristics. It runs like lightning and the data engine is fabulous - importing your data quickly and making intelligent guesses about how it's structured. You need ten minutes guidance to get used to how the dashboard front ends work, but once you've got it, then Qlikview is very simple to use.

Admittedly there are a lot of options for customising views and you get to them via large, complex options windows, rather than by a quick right click on the thing that you want to change, but fairly quickly you get used to where to find what you want. The positive side of all of these options is that you can create virtually any view that you might need.

Qlikview was a runner up though, based on cost. Actually it might have been dead last based on cost, because I never managed to find out how much it is. One of my criteria for software was transparent pricing and Qlikview's reps are a nightmare to talk to (I've had a few friends in IT back up that opinion too.) They want meetings; they want to know all about your business; they promise free copies of the software. What they absolutely will not do is give you a figure for how much it's going to cost to deploy the software onto x analysts' desktops and allow them to publish to a server.

It's a shame, but Qlikview was rejected on those grounds, even though I loved the demo.

------------------------------------------------------------------------------------------------------
Edit (22-1-13): Qlikview has moved to transparent pricing since this post was written (pointed out in comments below.) That's great to hear. Prices here.

Initial impression? The product mix looks a little over-complicated and pricier than Tableau. I'm not sure if Qlikview is targeting larger, more complex deployments than we've built, but I wouldn't change the conclusions below, based on anything I've seen over the past year or so.
------------------------------------------------------------------------------------------------------

Which brings me onto our winner. Tableau. Of course it's Tableau.

Tableau seems to be a well kept secret for now among analysts who've done their homework, but I can't see  it staying that way for long. It's a joy to use - simple, fast and powerful.

The software's actually quite restrictive in the way that it will allow you to present data, but that's a very good thing. It's like a best-practice data visualisation engine. No you can't draw 3D pie charts. Dashboard creation is simple drag and drop and the best way I can describe it is as if somebody took Excel Pivot Tables, made them nice to use, made them look good and then gave you the ability to publish them to the web.

Pricing is transparent (and very reasonable.) The desktop dashboard builder software has a simple cost per user. Not an ongoing licence fee, but a one-off cost, although you can pay for ongoing support if you want to. The server also has a one-off cost per user and that's it. No OEM nonsense and easy to work out pricing for scale, buy it once and do what you want with it. Perfect.

Other dashboard vendors really need to take a leaf out of Tableau's book on pricing. It's so much easier to pitch the idea to management of investing in software when you've got a clear idea yourself of exactly what it will cost over the next three years. If the dashboards you build are a runaway success then how much is an extra five desktop licences and twenty server logins going to be?

If you haven't tried Tableau yet, then give it a go. There's a free version for bloggers that's great for trying out some data vis and publishing it to the web. I feel like I'm giving away a trade secret here, but what the hell, you're going to hear about it from somewhere soon anyway. Tableau has very quickly become an essential analysis tool for me alongside Excel and if you try it, I'm willing to bet that it will be for you too.