Tuesday, 29 March 2011

The end of the ad-supported model

I've written once before about the limits of the ad-supported model. There's a finite pot of advertising money out there and at some point you have to find a different way to fund businesses on the web, or simply stop developing. The alternative is that eventually the whole world works in a media agency and earns money by advertising to itself, which is ridiculous and a scary image to contemplate.

Maybe that's why websites aren't called businesses very often; as opposed to being called social networks, or music discovery sites. Calling them businesses brings to mind difficult questions about funding and cash-flow. Amazon is called a business all the time though. Funny that.

The market likes to pretend that you can have large numbers of purely advertising funded web-based businesses and it will work fine. It's a collossal pump and dump scheme, which it's in almost everybody's interests to perpetuate, particularly banks like Goldman Sachs.

Wait a sec, you see that link on Goldman Sachs? The one you just skipped straight past? Open it now and bookmark it to read later (once you've finished this post, obviously.) It's a stunning piece of journalism.

Bookmarked? Good. On we go.

New startups like Color - who have raised $41m of investment capital - need to make the case that they can fund themselves using advertising as it's the only real revenue model open to them. Subscription services would take too long to pay back (if anyone ever subscribes at all) and the business plan would be immediately shot to pieces.

So you point at the big pot of web advertising money - £4bn last year in the UK, say new figures out today - and you boldly state 'we can win our share of that pot'.

It's a Dragon's Den reject's profit forecast.

"The market is £4bn and if we can just win 1% of that, then we'll all be millionaires and have super-yachts that make Roman Abramovich's boat look like a mirror dinghy."

"So how many Skyhooks did you actually sell last year?"

"Er.... [nervous shuffle] Have I mentioned the overall market is £4bn?"

It's in the interests of the start-ups themselves, banks, media agencies and bored marketing directors to pretend that the latest shiny web start-up will definitely work and that profit is just around the corner. It's even in the interests of the investors as long as they're not left holding any shares when the music stops. The game is to get out just in time and make your money on the way up. Like I said earlier, it's classic pump and dump but with a web 2.0 (aren't we on 2.1 yet?) twist.

I'm going to look on the bright side of human nature and assume that this can't carry on, because it's ridiculous. The emperor is parading around naked and everybody knows it, including the emperor himself. I know pyramid selling schemes never go away, but they don't usually take in whole industries for years at a time.

Eventually we'll have to recognise that worthwhile products cost more money than advertising can provide, although advertising is useful to keep the price down. Newspapers knew it for years, forgot and are just starting to remember again.

There is a relatively easy solution though and it's where I'm starting to think that we'll end up.

A large part of the cost base for internet businesses is in hosting and back-end IT infrastructure. Take Color again. Their app is very basic, whatever marketing spin the management might want to put on it. Never mind forty sensors, I'd be willing to be that 99% of the time a geographic fix to the nearest fifty metres or so would be fine. They'd probably be better suited by at least turning off the microphone, as capturing sound with the pictures is going to scare people off.

The cost (unless they're crazy) is in providing a photo hosting and delivery service to the entire world from day one.

Remember back when the internet was young and we were all going to have a website? In the end only geeks and businesses did it. Even this datamonkey doesn't have one. I've got a blog, sure, but I don't pay a tenner a month for my own site.

What we need is a model where you pay £10 (or whatever) a month for a virtual PC. A PC in the clouds if you like. Now I've heard that concept before somewhere...

Eventually, the cloud isn't going be about uploading your data, it's going to be about installing software. You host your photos, your videos and whatever else you like. You install apps, just like you do on your phone and they work with what you've uploaded (or help you to upload more.)

Crucially, you pay for the bandwidth and you only a little bit because you don't use that much. One payment runs all of the ad-supported loss making stuff you currently consume on the web. It's unlikely to be an extra payment either - you'll get it with your net connection. Even your mobile net connection. There's no risk in trying an app like Color, because you install it for free on a space that you already have.

When you think about that financial model, vs. Facebook on its own trying to pay for photo and video delivery for 500m active users, just by advertising to them, then all of a sudden the ad-funded model sounds like lunacy.

Think it won't work? How about an on-demand music and video download site offering DVD quality video and full album downloads that smashed the 12m user barrier back in 2008? It carries vast amounts of user-uploaded content and serves it on-demand. It has no venture capital funding at all.

Say hello to The Pirate Bay.

Forget the piracy issue, you're paying for the hosting and bandwidth with your broadband subscription. Social networking is very difficult on the bittorent model (unfortunately for Diaspora) because only you have got a copy of your photos and sometimes you turn off your PC. You need to subscribe to a service that provides you with an always-on connection.

I think this is where the cloud is going and the sooner we get there, the better, because this ad-funded hype driven pyramid scheme is getting seriously out of hand.

Thursday, 24 March 2011

Color me impressed

There's a new social network out today, called Color. I'd prefer colour, but moving swiftly on...

You take photos with your cameraphone and upload them to Color. It then shows you a photostream of everything that's been uploaded in your immediate vicinity. I like this idea. I like it alot.

Say you 're out at a venue for a gig. You can see a stream of pictures that other people are uploading, but the bit that I really like is if you visit the venue a week later, that gig is still captured for you.

Imagine wandering around a city and being fed a photstream of how where you are right now looked yesterday, or last week, or last year. That's an interesting idea. Visit Trafalgar Square and while you're there you could see how it looks full of protesters, or in the snow, or after dark. Bits of any big city are completely different depending on the time of day and day of the week and seeing that while you're visiting would be fun.

Like every social network, all it needs right now is for some other people to sign up.

And like most social networks, it's probably best not to raise the sticky question of how (aside from building a free user-base and then selling out to Facebook) it's ever going to make any money.

I'd also love to know who picked the name. Their SEO is going to be fun.

But it's cool. And that will do for today.

Tuesday, 22 March 2011

There are bad charts...

And then there's this one, which I've just been sent this morning. It was titled "The rise of social networks"

I've removed the creator's logo and a footnote that said "Source: UKOM April 2010", but other than that it's unmolested. I don't even know where to start with what's wrong with it. Anybody got an emergency copy of Tufte?

Monday, 21 March 2011

Dashboard & visualisation software: Excel Services review

It’s been a short wait for this first dashboard software review because… well because I’ve been busy building dashboards. Now with a long train journey from London to Leeds, it’s time to take a quick look at Excel Services. I wanted to look at this one first, because it doesn't seem to have been widely adopted and I've come across a few people who build very good Excel dashboards but who don't even know it exists. Excel is still the go-to solution for just about everything, so what about its Services sibling?

When I first came across Excel Services, I have to say my immediate reaction was “what took them so long?” Lots of people, including me, build dashboards in Excel - from simple report tables to large macro driven data explorers – and it’s a logical step from mailing them around a business to publishing them for viewing on a website: One version of the truth, no versioning problems and easier to read for people who aren’t comfortable in Excel.

Excel services sits on top of SharePoint and lets you publish spreadsheets from Excel 2007 onwards (*.xlsx files) to the web. You can then view elements of the spreadsheet in your web browser.

(microsoft don't really do pretty)

First the good, and publishing is fairly easy. You specify the sheets and named ranges that you want to make available and then from Excel, you save the workbook to the server. There’s not much more to it than that. Hopefully you already know how to make a reasonably good looking spreadsheet, so now you can make a reasonably good looking online dashboard.
Live data comes from data connections to pivot tables that refresh automatically when a user visits the dashboard. Not the fastest, but it works.

There are a few things you can’t publish and we’ll come back to the limitations of Excel Services later, as you’ll encounter some of the issues pretty early on. Big things that you can’t use are:
  • Macros. Your workbook has to work with worksheet formulas only.
  • Any kind of drawing object (including buttons and pictures) so you’ll have to make it look good by colouring in cells and giving them borders.

Once you’ve published the workbook, it’s immediately viewable online, but you’re going to want to build it into a proper dashboard with menus for the different screens and some element of user control. For that you use the dashboard builder and here’s where things get a little less smooth. The builder is fine and it works, but it’s not friendly. Creating filters and dashboard screens is very manual, pretty clunky and to be honest a bit of a pain. There’s a weird distinction between different ways of publishing and refreshing items on the server and you’re never sure which one your latest change needs, so you end up always doing all of them.

There’s also the potential to get into a hell of a mess with the builder if you don’t keep your files and working practices organised. One (local) dashboard builder file can contain the elements of several different online dashboards. Get two or three people building and uploading without remembering to refresh from the server and you can find your painstakingly created view has disappeared when somebody else ran an upload. It’s not that hard to avoid but it shouldn’t be possible at all.

For an overall idea of feel, have a crack at the skydrive. They’re not too similar in function, but they both have that hint of “we’re Microsoft, we know what’s best and if you do it wrong or can’t find a setting, then it’s user error”. They're also both slow. You know what I mean - Hotmail’s the same.

User interaction comes from setting up drop-downs on the dashboard and linking them to named cells in your workbook. You populate what’s in the drop-down from a workbook or database connection.

One major - and I mean really major – limitation is that you can’t make the drop-downs drill. So if you’ve got a list of countries in one and a list of cities in another, you can’t only show the UK cities if the viewer selects UK. This one issue alone kills Excel Services for serious dashboard work for me. As soon as you’ve got any amount of data, you have to be able to drill the options or there will be hundreds of items in a list. There are workarounds involving tree based lists but they’re not very satisfactory.

While we’re doing the negatives, here are a few more:
  • If you’ve got something that’s not allowed for publishing in your workbook, like a cell with a comment in it, you get no help at all about what’s wrong and have to go over everything with a fine toothed comb until you work it out. One of mine had broken links in a hidden named range (it had disappeared out of the names list – a ‘known issue’ apparently…) and that took ages to solve. There’s a download tool from MS to help, but frankly, it’s rubbish.

  • Viewers can’t download data from the dashboard unless they’ve got Excel 2007 or better. If you’ve got users on Excel 2003, my recommendation is that you don’t touch Excel Services with a bargepole. Getting the data out for viewers who are using 2003 becomes a complete pain.

Would I buy it? Well we did at EMI and it was useful. Talk to your IT people and if they’ve invested in the top-end version of SharePoint then you may find you’ve already got it.
For small dashboard tasks, it’s not bad and for my current needs, if we already had the licences then I’d use it to help with a lot of them.

Unfortunately though, we don’t already have the licences and it will cost somewhere around £8k to get them. With the other tools on the market (some of which we’ll come to) that’s too much when it only looks ok and is only ok to use.

Microsoft are being left behind in dashboarding and visualisation. In Excel, they’ve got what is still everybody’s first-stop BI tool, but they just haven’t made creation of an online Excel dashbard easy enough and they haven’t implemented the macros that make Excel dashboards so flexible.

In short, Excel Services is ok, but only ok. If you’ve got it already, give it a try and if you only want to publish small dashboards, have a look at it. For everyone else though, you can do better.

Friday, 11 March 2011

Dashboard & visualisation software: The criteria

Edit: If you're looking for the conclusions of this post, please go here.

I'm coming to the end of an evaluation process for dashboarding software and it's been an interesting journey. Some pieces of software I've loved, others not so much and a theme running through the whole experience has been a serious difficulty with finding information on the pros, cons and pricing of each platform.

I had a rant a while back about how dashboarding software is sold. Coming out of that post, some discussion on linkedin and some very helpful people who got in touch, I've finally got a reasonable picture of what's out there and roughly what the different options cost. Capturing that sort of knowledge is what blog posts are for, right?

Over a few posts, I'll line up the candidates, with reviews and the path to the platform that's come out on top. First though, with so many platforms out there, what was I looking for?

I do econometrics (marketing return on investment analysis) at Brilliant Media, a mid-sized, independent media agency with three offices and 100+ employees. We've got a lot of basic marketing response data - click through rates and monthly spends, that sort of thing - plus more advanced modelling and forecasts coming out of the econometrics. There aren't vast volumes of data (a few million rows max) and it needs to be visualised clearly for our clients, who are marketing directors and managers who may or may not be comfortable with maths, stats and charts. I'm not expecting a dashboard to work miracles, but to be functional, effective and look good.

These criteria are probably fairly typical for a smaller dashboard project where the output will be viewed internally and sold to clients.

  1. Off the shelf, not custom.
    I'm not interested in starting a major IT project here, I want something that works out of the box as quickly as possible and can be built and maintained by (SQL capable) analysts with minimal programmer involvement.
  2. Low cost.
    Always! Our clients aren't huge and dashboard server solutions for ftse 100 companies really aren't happening. We need to get off the ground for a few grand, max.
  3. Web facing (or at least has the potential to be.)
    We're mailing spreadsheets at the moment and need to be able to step to web-based delivery.
  4. Better looking than I can do myself in Excel.Which actually sets the bar quite high - Excel's not bad if you put some effort in.
  5. At least as flexible as Excel.
  6. Easier than building and then mailing spreadsheets to clients.
  7. Data stored on our own servers
    There are some quite tempting cloud based solutions emerging, but we need the capability to include confidential data, so they're out from the start.
Finally, a word on OEM licences. OEM stands for Original Equipment Manufacturer and it's a world of confusion in terms of pricing. If you want to build a dashboard that you're going to sell to somebody else then you need an OEM licence, if you're building just for your own internal business, then you don't.

The trouble is, OEM licence prices (usually) depend on the scale of your deployment. If you want hundreds of people to be able to log into your dashboard, the software will cost a lot more than if you only want ten, even though it's the same software running on the same server in the same building. That's annoying and it can become absolutely infuriating when you don't know yet how big the market is for your new product.

If you want a giggle, ask your IT department to find out how much Excel Services will cost to deploy externally on your Sharepoint extranet. You'd think it being a recognised tool from a big company like Microsoft and the fact you've already got Excel would make it simple wouldn't you? Give it a try.

So a last criteria...

Transparent pricing. If I still don't know what the year one cost for the software will be after a ten minute conversation with a salesman then, forget it.

Excel services will be up first. Check back for more!

Wednesday, 9 March 2011

Mobile web pages are better than their big brothers

Wallpapering fog had a face-lift recently and I hope the new look's an improvement. As well as making it prettier, I wanted to soften the overall feel and make the post text easier to read. It should look better on a mobile now too (Google Analytics says some of you lot read it on your iPhones) and mobile web page design got me thinking.

Having your capacity to fill a screen with garbage taken away is often helpful. Look at the difference between Facebook and Myspace; on Facebook, users can't turn the page into a customised HTML car crash and that's a very good thing. Although Facebook does seem to be gradually cluttering up with more options and bits of menus moving about - they must have been talking to their new buddies at Microsoft.

In order to work, mobile web and app pages are forced to follow a couple of simple design principles.

- A page fits on a screen (horizontally at least, vertical scrolling is ok.)
- Crucially, a page has one purpose. There isn't room for more.

The unintended consequence of this screen real estate discipline is that publications' websites can be easier to read on a smart phone than they are on a 19" widescreen monitor.

Take The Guardian. It's got a fairly typical online newspaper look, but there's a lot competing for your attention on this homepage.

The stories that I came here for are relegated to the left-hand column, below loads of menus, beside a special feature and jostling for space with links to jobs and online dating.

It's the same picture when you click through to a story. I've still got a list of headlines on the right-hand side even though I've just seen a list of story headlines on the homepage and I picked this one. I doubt that even one in a thousand people use the find your MP form.

And just for fun, let's turn the ad-blocker off too. What did I come to this page for again? I've forgotten. The side-bar ad pushes the MP form and headlines lower down the page, but they're still there.

What I actually came here for is the left hand half of the page. All that other clutter is a distraction.

This is the Guardian Anywhere Android app, (see my Android App essentials)

The front page is a list of categories. You click on a category for a list of stories. Stories are a single picture, above clean, well formatted text.

If you want the menus again, you click the back button. That's it and it's beautifully simple.

On my HTC Desire HD, which has got a fairly large screen, I actually prefer reading The Guardian with the app than reading it on a monitor. I know there are major problems with minimalist design - starting with the fact it's very hard to make any revenue if you can't have ads - but as screens get larger, publishers are making web pages more complex and that's a bad thing. They could learn a lot by looking at their own app screens and then wielding a digital scalpel.

Monday, 7 March 2011

Too clever by half... how to over-complicate your marketing.

When I was working at EMI, a hot topic was the single pre-release window and whether it works. Sony and Universal have recently decided that it doesn't.

Pre-release involves letting radio stations have a single before it actually goes on sale. In marketing terms it sounds like madness - you play radio listeners your track for a fortnight, giving them time to get interested in it, record it, pirate it and potentially even get tired of it before you ever try to sell them a copy.

There is (or was) a logic to pre-release. Chart position is (was) an incredibly strong form of marketing that leads in turn to more radio airplay, so you build up as much interest as possible over the pre-release period and trigger it all in a single week. Then you cross your fingers for a number one hit and a snowball effect as that chart position leads to more and more airplay and an extended chart run.

My gratuitious bracket use is a hint to the problem with this strategy. Twenty years ago it was probably quite sensible - if a little too clever for its own good - but fast forward to 2011 and it's a disaster.
  • Regardless of whether the label says a single is on sale or not, if the album's out then you can buy the track on iTunes. Do you really want the b-side that much? Come to that, what's a b-side in 2011?

    In 2011, you've just got a strategy that makes a mess of your marketing timing between iTunes and Radio.

  • If you play people music they like and tell them they can't have it, what do you think they're going to do? Wait patiently for two weeks and then rush to HMV, or pirate it right now? No prizes for the right answer, it's too easy.
It's unbelievable that inertia in the music business has meant that it's taken this long to change strategy.

At least that one was probably a sensible strategy once. I was reminded of it this morning by the news that Heinz are going to pre-release a ketchup flavour on Facebook.

Why on earth would you do that?

Apparently they expect the 3000 bottle run to sell out in days and it probably will, but then what have you achieved?

Let's assume a best case. Facebook is a well targeted advertising channel for the housewives (that's media targeting housewives - they don't have to be women) who are typically responsible for ketchup shopping. Yeah, right, but moving swiftly on...

You generate loads of interest in the page and shift your 3000 bottles in less than 24 hours. We're obviously pretending here that most of the PR coverage isn't going to be the marketing industry taking to itself.

You've got 3000 happy Facebookers and lots more who couldn't try your tasty new condiment but really, really want to.

Now they've got a few weeks to forget all about ketchup before they visit Sainsbury's. Perfect.

As a strategy, it's over-complicated and inefficient. Pre-release music was an example of an industry obsessed with the music charts assuming that everybody else was still obsessed with them too.

Ketchup on Facebook sounds like an ad industry that is obsessed with Facebook assuming that everybody else is just as obsessed as them and will want to buy everything there.

Friday, 4 March 2011

Germany's iPad fanboys are the most dedicated

There's a new iPad out. What do you mean you hadn't noticed?

I've been playing with Tableau's visualisation software and thought it might be fun to have a look at which countries have shown the most interest in this new fondle slab.

So, hit Google Trends, take the level of searches for "iPad" in 2011 as a base and compare the level of searches for either "iPad2" or "iPad 2" over the same period.

It's not very scientific I'll admit (it uses Google Trends data for a start) but here's a view of how much interest the new tablet has generated this year vs. the generic name

And a zoom into Europe