What should come first, critical mass or a business model?

Reid Hoffman, chairman and president of products at LinkedIn, recently said at a conference that web service companies should concentrate on building a critical mass of users before they worry about how to make money (I assume he’s referring to services where the concept of a ‘critical mass of users’ is important – such as those with a strong element of social networking and user generated content for example).

There are lots of successful companies that did not monetize their users until a long way down the line. Just look at the biggies: Google, YouTube, Facebook, Skype, LinkedIn, MySpace … and what is their primary business model? Advertising and premium services. Yet so many VCs keep telling me that the business model has to be nailed from the start, and by the way "advertising is not a business model." From what I can see, advertising is making lots of people very, very rich indeed. Advertising on the internet and mobile is exploding. Advertising certainly is a business model and a very lucrative one at that.

If you have the critical mass of users. Which is starting to sound like bubble1.0 again: then it was about getting as many visitors as possible to your portal, now it’s about getting as many members as possible in your social network. It’s an improvement this time round, though, because users will be more engaged and you know intimate details about them such as the colour of their underpants. And a big company that already knows that this sort of thing works might buy you.

But since you’re not one of the first web2.0 companies that could attract lots of new members at a low cost because it was all so new and cool and there wasn’t much competition, there’s a tough challenge ahead: you need to expend lots of resources in the early stages to build critical mass with precious little return. To have the best chance, you need to develop partnerships that can enhance the service and bring more reach, have the resources to develop your offering faster than someone else can copy it, and make sure you’re in the right place at the right time. It’s a crowded market, and few will survive long enough to reach the tipping point.

Opaquevertising

In mobile and internet advertising, if you want to know how well your campaigns are working and be able to fine tune them to perfection, what do you need to measure?

You need to take a whole-process view from ad wording to country/channel/device placement to impressions to clicks to visits to registrations to loyal returning users. And to track this you need to put in place measures that link your mobile advertising agents' statistics to your own server-side stats - no mean task, given that the way they break down the information isn't the same as yours.

Mobile advertising agents typically provide some basic stats like CTR, CPM and CPC. But they don't provide enough breakdowns to enable you to properly measure the quality of their publisher network. Why should they? Their incentive is to sell you clicks, not customers. The value-added part is left to you sitting in your dark room with your poor server to help work out the rest.

The classic measures offered by most mobile advertising agencies are:
- How many people see the advertisement and for what price: Impressions / day & CPM (Cost per thousand views = Spend / 000 Impressions)
- What is the level of interest in what the advert is promising: CTR (Click-through-rate = clicks / impressions)?
- How much does it cost on average to bring a visitor to your site: CPC (Spend / Clicks)

Unfortunately these measures are practically useless without an appreciation for the impact on your registered - and more importantly, active - user base. So you need to know:
- What is the registration rate for your service? Registrations / Visitors per day
- How much does it cost per acquired registration? CPA (Cost per Acquisition)
- How much does it cost per returning loyal user? CPLU (Cost per Loyal User)

And if you want to know how best to target your advertising you need to measure this whole-process 'funnel' (Impressions/day, CPM, CTR, CPC, Registration Rate, CPA, CPLU) in terms of - ie broken down by - the advert wording, presentation method (banner or text or in-game insert…), timing (Sat, Sun or weekday…), publication channel (portals, communities, search, games, downloads, adult...), target country, target manufacturer / device / platform, visitor demographics, and so on …

No mobile advertising agent gives you enough information to work this stuff out. So you need to be clever in how you set up your advertising campaigns with these companies. Try out different combinations of ad wording, country / channel / device targeting and track the results. Find ways to link server visit stats to the agencies’ click stats and give them a hard time when the figures don’t stack up.

And really make sure that server-side you’re doing your best to track the source and behaviour of your visitors, registrations and returning users. It will pay off in the long term.

Madvertising

Our first mobile ad campaign generated over 1000 fraudulent clicks and flushed our daily budget down the pan in under 2 hours. Good job it was just a test run on a 1 cent CPC for a max 10 dollars spending limit. [update: it turns out that only about 100 of the clicks were fraudulent; the agency we were using had an unexpectedly high publisher inventory in india which even at the lowest possible CPC rate gave thirty thousand impressions in 2 hours. We were not actually charged for fraud clicks. They were registered by our server but were not counted in the agency's spending reports] The prime suspects? India blatantly in first place followed in a distant second by an unabashed South Africa. All the other english speaking countries were either asleep in their GMT +/-8 time zones or weren't paying hundreds of workers to speed-click on adverts for a dollar a day. It clearly pays to be cautious when starting out with this high risk advertising medium. We now need to do more extensive tests. With global advertising everyone is in different time zones so we need to run each campaign for at least 48 hours to be able to start making comparisons. That way each country will have 'experienced' the adverts for a full 24 hours during the same day of the week.

Two types of buzz

Whilst talking to a friend of mine who works for Come&Stay it was clear why we should pay more to attract mobile visitors than internet visitors. It's because they're already mobile web users looking for the sort of service we provide ... on a mobile. And that also reinforces my belief that our early adopters are not the people who get goosebumps using Twitter. They are not people wanting to broadcast their trazes on the internet for other web2.0 enthusiasts, wannabes and curious surfers. They are people who don't know or don't care about web2.0 or mobile2.0 but who do like exploring their mobiles to do fun, cool and yes, even useful things. It's interesting how much time companies spend trying to create buzz in this techno ecosystem that does not reach the early adopters they are really targeting. Of course they want to attract the attention of bloggers, companies and investors. But it's better to win over real customers first.

KPI

We spent a couple of hours today reviewing the launch plan and preparing for tomorrow's board meeting. One thing that investors want to see is KPIs. We need to show that lots of people are using our service, that they're not all signing up once never to return again but are actually using it, and that their numbers are stacking up by the bucketload. Doesn’t that sound familiar? Reminds me of bubble1.0 where the volume of visitors to your website was considered the holy grail. This time round, the red herring is the number of active members of your online community. There are now hundreds of me-too start-ups without a business plan or a business model hoping to wave as many adverts at as many people as possible whilst they whistle through their ‘community’ from registration to emergency exit stopping only to look at a few pretty boys’ and girls’ profile pics along the way. Their visitors are all frequenters of their competitors’ communities too. If their offering is not something unique that their customers would happily pay for if it didn't just so happen to be free then they're facing an uphill struggle all the way.

A recent article on Two Point Oh on how to create a thriving web2.0 business listed 10 tactics for success:
#1. Create a sense of community
#2. Have a simple proposition
#3. Ensure viral growth
#4. Do continuous R&D
#5. Build a platform for advertisers
#6. Let the outside in and the inside out
#7. Own the audience
#8. Don’t forget to be profitable
#9. Get your content mix right
#10. Sell something

In my opinion the last point is the most important of all: Sell something. If you can't figure out what you’re selling to your customers you don't have a business - you have an expensive hobby.

So another KPI for us will be: how many of our customers are buying our premium service and how many are just clicking on adverts on their way out the door?

Break it Down

We're planning many methods of publicising our new product that won't break the bank. What are some of the options?

- Mobile advertising. Still in its infancy and possible to get a good deal to reach users either globally or in specific countries or by specific phone / model / manufacturer / publisher type. Lots of companies popping up all over the place. Some examples: Admob, Admoda, DeckTrade, Google Mobile, ...
- Internet advertising. We get thousands of visitors to our website, and thousands of downloads of our bluetooth software per day...
- Existing users. They deserve to be at the top of the list, they like our Bluetooth product and we generate revenues from this. We have a huge email list and an active forum as a platform.
- Blogging and other bloggers. Time to get the referral services fired up: Technorati WTF ('Where's the Fire?'), Reddit, Digg (too damned difficult to get seen now!), StumbleUpon, Ping-o-Matic, del.icio.us ...
- Partnerships and alliances. Of multiple types. Content swapping, API 1-/2-way swapping, cross-promotion, joint development, client cross-referral, ...
- Events. The most expensive way of marketing but highly targeted at specific audiences. Best done in conjunction with partnerships to get a good deal.
- Viral spread. Give users a way to invite friends and even strangers to join them in using the service.

This list is not exhaustive of course but the devil is always in the detail. You have to understand the potential reach versus cost of each approach broken down by various factors such as geography, phone model, age, usage etc.

I can feel a very large spreadsheet coming on...