Pinstorm CEO Mahesh Murthy shares his views on Twitter’s ad model in his WSJ column.

We’ve been waiting years for it – something that would let Twitter justify its billion-dollar valuation.

While the company already is profitable, mostly from licensing fees paid by Google and others for the right to search and index their tweets – that’s pretty much a flat-line model. Like what Google had before it stole the Overture idea and called it AdWords.

What we’ve just seen is the first hope for Twitter’s hockey-stick revenue curve – their advertising model. For those who care, it’s a little like Google’s, with some differences. You’re starting to see one “Promoted Tweet” result on the top of Twitter search results that some brands are paying for.

Interestingly, the promoted tweets are still all-text, and just regular tweets that the brands are already doing to their current followers for free – with just a raised profile so everyone who searches for a particular term – and not just one’s followers – can see it.

I have two sets of observations to make – one relating to the obvious comparison with Google. And the other relating to a bigger change I think advertising will go through as a result of all this.

Let’s get the easy stuff out of the way first.

Will Twitter’s revenue stream be anything as big as Google’s? In a word, no.

Here’s why.

Very little of Twitter usage is search. If you look here: you’ll see that Twitter has somewhere over 80 million who visit it in a month. But look closer and you’ll see that less than 3 million use Twitter search. Compare that with Google’s 300 million+ people who search and you’ll see a difference by a factor of 100.

For each of those searches, Google shows up to 30 ads. Twitter says they’ll show just one. Even if Google averages 8 ads a search, that’s 8 times the number and revenue Twitter can show on a search-for-search basis.

Google charges per click and gets about 25 cents for each one on average. Hence 1,000 searches generate 100 clicks or $25. In yield terms this is a CPM (cost per mille, or thousand) of $25. Twitter will probably have an impression-based model. Typical impression rates that clients might pay for a text-only ad are likelier to be lower – around $5 or $10. So ad-for-ad, revenues are likely to be much lower.

Back-of-the-envelope calculations say that Twitter’s 3 million searchers – and their 20-odd million searches every month, even if fully sold at this rate, should generate less than $200,000 a month in revenues. Chump change compared to Google’s near $2 billion a month revenue run rate. Billion dollar valuation? We’ll need another trick for that, thank you very much.

Google has an automated ad-serving system that lets over 1.5 million small and medium-sized companies directly place their advertising. I have no idea if Twitter is planning to have one of its own – but it doesn’t have one at present. Another black mark. You can’t sell this door to door. Just getting a few Starbucks and Best Buys won’t power it to big revenues.

With all this, I don’t think Twitter can be a Google-killer in revenue-terms any time soon.  But there may be more jokes up Twitter’s sleeve – as this alone won’t move the needle much.

Now to the more interesting bit.

How will this change the advertising and marketing business?

  1. If you’re a marketer, and if you don’t regularly tweet, you’re toast. There’s over a 100 million people who already do – and believe me, you’re already part of their conversation. And it’ll go on without your intervention and moderation if you don’t join in. You need to set up an internal infrastructure to do this – or get a firm on the outside to take it on.
  2. Start tweeting – and get ready to face the music. You may just want to reach out and make prospects happy – but people will come to you with service complaints. Are you geared to handle it? At Pinstorm, a company I’m associated with that offers such services to various brands, one financial services brand alone caters to over 100,000 mentions a month, and tweets out about 100 times a day.
  3. It’s not a joke for a business to integrate their brand marketing, product marketing, customer service, investor relations and corporate communication folks – as the same Twitter account is expected to talk for your entire company. People on Twitter expect one face for your brand – you can’t shy away from a question saying “Oops can you address that to customer service, please?”  You have to go through some internal re-engineering to make this work.
  4. But make it work – and its magic. This is the most direct connect you can have with prospects and their perceptions. It’s real-time brand management in the true sense of the word. There are case studies already out there on how folks as diverse as ICICI Bank and Café Coffee Day use Twitter to connect with customers seven days a week. And in the process, bring about a big positive in brand perception.

Now if it’s difficult for marketers – it’s sheer murder for the established order of advertising firms.

  1. People tweet 24 hours a day, seven days a week. You’ve got to listen to all this – and reply in real time. And the reply can’t come from BPO service-deck folks. Trained copywriters know how to speak in a brand’s voice. Can you have your creatives give up their vodka-soaked weekends and work shifts that cover 24 x 7, because their brand’s customers do? You’ll have to.
  2. That’s just for tweeting. Can these folks then also decide how much you’ll pay for a promoted tweet – in real time?
  3. How will Martin Sorrell make his money when it’s not easy to do neat media buying deals through Group M in this business? All basic tweets are free – and commissions or management fees on brand-specific promoted tweets can’t quite make for a significant revenue stream, can they?
  4. It becomes more complicated because any marketer will ideally want its audience to become its followers – and hence not have to pay for tweets to them. But agencies will want this to not happen as the money is in talking to non-followers. Smart marketers may just pay for tweets that essentially say “Follow me” so that the cost of contact goes to zero once they get the millions they want to follow them. Who will win this conflict?
  5. Google’s ad model was the first big blow to the ad agency way of life – because it embraced pay-for-performance in an industry where agencies typically want fixed fees for their work whether it’s successful or not. Twitter’s model adds a second layer of complexity – real time, 24-by-7. Put these together – accountable real-time marketing – and you’ll have to burn the house to the ground and build it up again. The current 4-monolith structure of WPP + Omnicom + Interpublic + Publicis is all about unaccountable, 9-to-5 work – and hence not likely to survive this unscathed.
  6. The regular ad agency system committed its first slash of sepukku when it separated the media and creative departments into different companies. This may have been all right in the non-digital world where the former depended on media kickbacks and the latter on TV production kickbacks to make the real money.
  7. But this separation has made them particularly unsuited to the digital world: will you not do Twitter for brands because it’s free and hence you can’t earn any media commissions from it? Will you not do Facebook when you know users don’t like to click on ads there and leave the environment to come to your landing pages? If you could tenuously separate the medium from the message in traditional advertising and split them across different firms, you certainly can’t do that in digital, where in McLuhan’s terms, the medium is really the message – and vice versa.
  8. And all this isn’t just restricted to Twitter, Google and Facebook. With the rise of the iPad, YouTube and DTH, soon print and TV are also going to be fully digital. Fully trackable, fully measurable, fully accountable and fully interactive in real time, 24 by 7. There will soon not be anything that is separately digital. It’ll all be digital.
  9. Digital may be 10% of the world’s spends right now – but digital won’t get subsumed into the system of TV spends. It’ll be the other way around. TV and Print spends will start getting benchmarked to the norms that digital has already set. How many people actually viewed the commercial or the print ad? How many clicked away from it? How many said in real time they liked it? How many didn’t? How many talked about it? How much did the brand likeability score increase in the 10 minutes after the ad was seen? How much did sales go up the next day? What was the ROI of the print ad? When all of this data is going to be available, how will the world of flat-fee advertising survive?
  10. Through crisis comes reinvention. This change will allow a new generation of data-driven, planning-driven and idea-driven ad firms to surface. Firms who can quickly help their clients react in real time to take advantage of opportunities. Firms who will take some or all of their compensation based on the result of the success of the work they do.

It’s not just the agency world that’ll need to reinvent itself. The marketer will soon come to their computers to do a different sort of job.

It’s a great leveling opportunity. A smarter marketer who can take advantage of all this free viral media can easily outrun a big, rich marketer who simply throws a lot of money at TV, print and display ads.

Where once a firm like Proctor and Gamble had 4-year product life-cycles and you could spend a year shooting ads on Brazilian and Mauritian beaches, perfecting the communication in anticipation of launch, we now see products with a planned obsolescence of 4 months. Some firms like Google never actually even launch their products – preferring to have them perpetually in beta.

A marketer’s job is increasingly becoming like that of an air traffic controller. You come to your large-screen, see your campaigns in flight around the world, in real time, and you track their performances.

You make a few adjustments, raise the altitude of some that are dropping, and then you go home after your 8 or 12 hour day. And as you walk out your counterpart who does the other shift comes by and takes over your screen, piloting the campaigns through their next 12-hour period around the world.

It’s a scary thought. But a beautiful one.

I for one, look forward to living in it.

We created a questionnaire to find out what India thinks about insurance. We wanted to find out what your experiences and perceptions were with life insurance and general insurance companies.

We posted the survey on Twitter and Facebook and received 134 responses. Statistically, this sample size indicates that the answers here may have a 8% margin of error, at a 95% confidence level in the validity of the answers. We believe the responses are a fair indicator of the questions asked in the survey, and reasonably indicative of the opinions of people online in India.

As promised, here are the unedited results of the survey, 6 questions on what India thinks about insurance.

All numbers on the axes of the graphs are absolute response numbers. All numbers where given inside the graph represent relative shares of answers in that instance.

Click on the images for a larger graph:








The poll-takers here were self-selected and no claim is made to the data’s representativeness of the market as a whole. That said, we find through experience that the learnings from surveys with sizes over 50 are reasonably representative of the opinions of the broad target market, in this case, people online.

For more details, feel free to write to us at

Contextual targeting needs work.

The first attempts at the long-awaited ad-based revenue model for Twitter were finally sighted. We saw the first glimpses on our Twitter account, while accessing it from an Indian IP address. We’re not sure it’s a global-rollout yet and even these sightings have been stacatto – with the ads appearing and disappearing at random. We were lucky enough to catch a screenshot.

As many pundits guessed, the new ad model is a take on the Google AdWords model with a few changes. First, a little digging at the server side unearthed that these were being served from a machine called “ConTwext” – an obvious take on the idea of “context”. Other IPs we could back-resolve showed “CT1″, “CT2″ and such – these seem to be ad servers that the ConTwexts are being served from.

The ads are even shorter than the 95 character limit (25/35/35) that Google places, and they have no headlines. They use long URLs in contrast to the Twitter standard of shortening URLs.

Here’s our first take.

First, the positives

1. The ads are seamlessly integrated with the Tweets. They don’t stand out like an eye sore. This is a good thing. But how will it affect CTR?

2. There seemed to be a large and diverse inventory of advertisers as virtually every tweet we saw had a ConTwext ad attached to it. They have obviously started getting their ad sales act together.

Then the negatives.

1. The contextual relevance of the ads need a lot of work. We understand some targeting, but others are way, way off the mark.

2. We were unable to see any geo-targeting, which we understand will be a key killer feature. But it could be an early test and we’re awaiting that in the next roll-out.

Have any of you started seeing this yet?

Twitter Screenshot