How The Future Of Web Advertising Is Linked To The Call [Updated: It Already Is]

[The article below was written on the incorrect premise that Google is not able to link phone calls with Adwords. It already does, as we’ve covered in the past here and here. We regret the errors below, and we’re working on follow-up posts.]

We’re all now familiar with how adwords campaigns on Google work. You buy keywords commonly used in search terms, such as “plumber in X town”, and send people to a response mechanism, usually a web site. But increasingly that response mechanism is not a just a web site but a phone number as well – sometimes it’s even just a phone number. But these days it’s rarely an ordinary number – it’s usually a ‘smart number’ that performs certain kinds of actions and sends data, just like browser calls a web page and sends data from that page. These smart numbers can be made to grab an RSS feed, play a sound file, make the caller fill out a form with their voice – just about anything.

Increasingly we are seeing tech startups address what you do with that phone call and the data and analytics that can be pulled from it, just like on the Web.

While Google and Facebook look at this area with their pet own projects, startups have appeared on the market to address this, such as AdInsight, Tropo, Twilio and Iovox, among others.

And the news that AdInsight, raised $2.6m from Eden ventures recently threw into sharp relief how competitive this market is likely to become. What is at stake is a billion dollar market of phone calls, which the majority of a time lead to real business being booked – far more than that generated by web advertising.

So imagine that you sent Web users to a phone line, but when they called it the line was dead resulting in an obvious loss of business? You’d be pretty annoyed. TechCrunch has uncovered just such an example of this happening, with shocking results.

We understand that a recent UK Google AdWords campaign involved a “several million pound” AdWord campaign spend where over 80% of the phone numbers associated with the campaign were linked to dead phone lines. The problem was spotted for some time and in the end tool two weeks to fix – and though Google wasn’t to blame, the uncovering of this massive failure put a UK startup, Iovox, into the spotlight of a potential bidding war between Google, Microsoft’s Bing and telco giant BT. Here’s how the scandal of the £4 million campaign that virtually failed went down.

A company connected their Google Adwords campaign to a voice-response mechanism. Customers would call the numbers advertised and be put through to a call centre. To that end a service provider – which we understand to be a major continental European telco – was brought in to provide the phone lines for the AdWords campaign. They employed an off the shelf PBX system to terminate the calls. But our sources say the telco only put in a handful of phone lines, around 8, – and only one actually worked. The result was that if any call came in on any of those lines, the rest would simply be sent to dead air.

Although the startup won’t comment on the incident, TechCrunch understands that Iovox was brought in to assess the efficacy of the campaign, and it uncovered that only 19 percent of attempted calls were actually being picked up. The client went back to Google, and passed on Iovox’s findings.

The client eventually told Google the fault lay with the telco. We understand Google later called for a meeting with Iovox. But this incident has highlighted a startling conclusion.

Google itself literally has no way of tracking the efficacy of adwords campaigns associated with phone numbers. This is a big problem for the future and Google knows it. And we understand from sources that an internal pilot to create such a platform has been dumped after failing to work.


As much as we think the old world of phone numbers may one day disappear, and we’ll eventually be calling each other through Facebook or something, the practical reality is that they won’t.

So the interface between the Web and the PSTN phone network is not going away. Indeed, the relationship is likely to get richer and deeper, and this is being reflected in how online ads are responded to using phone numbers.

So, someone calls a normal phone number that terminates on a PSTN switch. That hits a flowchart on a server and based on an API or time of day they perform some action. That platform can then send the called to a call centre, or start an IVR process, more or less anything.

It could be sending the call, for instance, to Livebookings so that the caller can book a table.

The important thing to remember here is that you can also attach meta data to the call, so it’s almost as if an individual phone number becomes a web page not a dumb bunch of digits.

Different companies use this method for different things like tracking whether those calls are answered, how long they stay on the call, at what point they drop off etc.


Now, when a plumber who advertises though a phone number or other platforms like Google, Yellow Pages, newspapers, or flyers, or whatever, he or she just wants a pie chart which says which platform is doing best. Measuring that online is slightly easier – email signups for instance. But plumbers and most other SMEs want phone calls, because that leads to real business. And they don’t want to have to be tech savvy – they just want data which tells them whether the number on their van, business card, web site or whatever gets them their most business and then they want to double down on the best option.

This is where a company like Iovox or Adinsight comes in. They provide these individualised numbers to any business. But the problem is that even though we are increasingly using smartphones, numbers are still dominant and still dumb. And that won’t be fixed any time soon.


But, as I said, there is no ‘last mile’ solution built into Google to track marketing efficacy between phone numbers and Adwords.

No one works with Google to provide phone numbers. So companies are running ad campaigns on Google and getting phone numbers from other third party providers. Often these numbers are found in newspapers.

Since people calling these numbers usually declare that it’s an ad they saw in a newspaper an advertiser will use a keyword to trigger data capture. Thus, one customer of these third party phone number providers has 6,000 keywords. Yes, 6,000.

Some can be expensive, like “travel to the UK”. What the number providers do is take a pool of those keywords and assign a phone number to that pool, or they can assign individual keywords to individual phone numbers. Then if someone calls and uses one of those keywords you know you’ve got a sale and can track it.

So if you’re spending £500,000 on these keywords tracking this spend become pretty crucial.

When someone pays for a keywords they play per click (whenever that keyword triggers a sale).

But, what advertisers rarely realise is that once someone sees phone a number they usually save it to their mobile phone. That number can then be held in that person’s phone, almost forever, even long after a particular keyword campaign has finished.

The advertiser is no longer paying for the ad in the paper, but they are still paying for the ad with the number provider. Different companies treat the number differently. Iovox, for instance, keeps each number for as long as the client is playing for it and the number remains live, sending calls to the right place. AdInsight tends to rotate them in a different manner (see below).

So a single phone umber can have a life long beyond the advert.

And of course, that number can become very valuable. As Ryan Gallagher, Iovox CEO, told me: “We have a number associated with one newspaper advert that only ever appeared once in the paper nine months ago. On a Tuesday. This month there were 1,400 phone calls to that number.”

Can that number ever change and become associated with another advert? Technically yes, if the number is rotated. But one phone number can often generate enough revenue to keep it open in theory forever, if the sale resulting form a call to that number is big enough. Some industry estimates say rotating numbers could be responsible for 30-50% of phone leads being junked.

A plumber can pay £5-£15 a month for a phone number, often because it’ll make hundreds of pounds in a sale to them, while a lot of businesses will pay £50 a month for Google AdWords, just for visibility. But few businesses want clicks. They want phone calls or people to come and see them, because that makes a sale. That’s where SMEs are at.

Unfortunately, while Google and Bing do a great drop of tracking online behaviour, SMEs don’t give a shit, to be blunt – they care if someone calls them, not if they see a we page. That means that keeping a phone number open and trackable in terms of response can be worth far more than AdWords. Visibility online tends to be of far less concern to small businesses.

But Web traffic right now is hardly ever linked to someone walking through the door of a restaurant. That link can rarely be tracked, unless you link up a phone number with a Web site or campaign.

So what is the market worth? Just looking at the UK, there are 4 million businesses. Some 50% of business for SMEs comes through the phone. Each will have multiple phone numbers, so that could mean billions of pounds in potential sales.

Now, while Google has 65% of the search market in the UK against Yell, Bing, TouchLocal etc. search traffic for SMEs is now the key battleground and Google’s share of that market has been dropping as competition from other places, like Bing, increases.

That’s why the battle for SMEs is a key battleground, and why this whole business about linking the web with phone numbers has become so important.

It’s why Google is going against web directories everywhere. But the real problem is they don’t provide anything above and beyond what the traditional web-only directories provide.

So next they need to link offline behaviour with online. Android will be part of that strategy, and perhaps Near Field Communication. They’ll need things which link to the EPOS system in a restaurant or shop that registers that person coming in as a result of a web search.

Right now, LiveBookings allows people to bring in a code, then the restaurant knows that person booked online. LiveBookings is paid when that person physically turns up.

What’s at stake is literally analytics for phone calls. You need to be able to tell how many times a number is used, the lifespan of that number, how long people stay on the phone, at what point they they hang up in the call, if there’s a sale etc.

Smart number providers like Iovox and Adinsight, and tools providers like Tropo and Twilio treat these voice calls as data.

A phone number can also be linked to an ad so that if it’s redirected to a call centre they know what the person is calling about, creating a higher level of service and therefore probably a sale.


The tech startups in Europe and beyond, in this space, tend to divide into two, with some raising significant funding.

UK-based AdInsight connects a user’s browsing history with how he or she then interact with the company on a phone. AdInsight has TUI Travel, Thomas Cook, RAC and British Gas among its customers.

The company’s flagship product, AdInsight Clarity, links Javascript on a web page to a company’s site and individual ads, which then tries to link web behaviour with the eventually call to a company. It can also attach unique phone numbers to specific online ads and integrates Google’s AdSense and AdWords analytics into its reports.

There is also Tropo, which is more about tools for developers, allowing them to add voice, SMS, Twitter and IM to applications. It’s effectively an application platform that enables web developers to write communication applications – such as deploying voice and telephony apps – in familiar languages such as Ruby, PHP or Python rather than the more obscure VoiceXML.

Twilio, (which has raised $36m in funding), makes voice and other telephony APIs used by developers in web and mobile apps. It lets developers incorporate calling functions directly on to sites. The company’s voice API, which lets users make and receive calls through those apps, recently expanded to 10 European countries as well as the U.S. and Canada.

Microsoft also recently offered Twilio voice and text APIs to developers on its Azure cloud platform. Twilio effectively offers developers similar tools to those that telecoms carriers want to sell, such as the BlueVia platform touted by Telefonica. Unfortunately, Twilio will need a lot more countries to be ubiquitous.

Finally, local UK/European player Iovox (which has raised a mere $700,000 from angels) has a priority solution which is closer to AdInsight’s but which concentrates more on making the actual phone numbers people call ‘smart.’

AdInsight tracks the users and rotates phone numbers, but it does not assign phone numbers to adverts. Instead it assigns the numbers to website visitors. For this reason all phone numbers for a client always belong to that client and are rotated automatically once visitors are no longer on the clients website. This allows AdInsight to report on the path of the website visitor before, during and after the phone call, including how they found the website, what adverts they clicked on and what keywords they used, as CEO Ross Fobian tells me. By contrast Iovox’s solution doesn’t rotate numbers associated with campaigns.

Twilio and Tropo provide tools to developers. Iovox is more of a turnkey approach. Do thousands of developers beat a turnkey solution like Iovox or Adinsight? Maybe, but it’s like Linux versus Microsoft – companies that want this kind of voice analytics tend to want someone to heavy lifting.

A huge directory business like – France’s Yellow Pages – are unlikely to deal with a lone developer who has built something on top of Twilio which has no aggregate data, no analytics, and no experience. Do companies that require these voice services want to be liable for some random developer, and get sued for it? On the other hand, Twilio’s view is that thousands of developers end up trumping the likes of Iovox.

Where Iovox stands out is that is doesn’t just provide information on a customer’s account, but on aggregate data as well. That means a customer running adword campaigns associated with properties can get data in aggregate on all of Iovox’s property clients. Its clients include Microsoft Bing, BT, Yellow Pages (Ireland, Belgium) Live Bookings, Zoopla among others. In Europe they are growing fast but looking to the US.


It’s clear that Google is going to enter this market, though we understand their own in-house efforts are not succeeding. At the same time Facebook just stared doing a pilot in the US, trying to track calls to SMEs from its social network. So the question is, would they pick up a player in the market to help them?

Google could acquire Twilio, but they would be acquiring tools for developers, not customers. Acquiring Iovox would mean using it to upsell to their existing customers and help them grow into SMEs. But If someone like Bing was to jump into the market with Iovox or AdInsight, they would acquire those customers as well as a solution, propelling them a lot further down the road.

It’s that last hurdle, that ‘last mile’ of the phone call that Google, Bing and Facebook ultimately want to bridge. (Yes, there are technical solutions liks P2P (Skype etc) and the newer RTC Web / Web RTC (Real Time Collaboration on the World Wide Web) effort. But they are years and years away from replacing PSTNs).

And it is a big prize – linking the Internet with every person who can receive a phone call on the planet today.

Update: Google has come back to us with this response:

In the article, it is stated that “Google itself literally has no way of tracking the efficacy of adwords campaigns associated with phone numbers.”

In fact, this isn’t correct – both Click to call ads, Call Metrics and bid per call enable this.  Moreover, for some of the reasons cited in the piece, these have been successful products: Click to call ads generate over 10M calls per month on mobile, and Google connected more than 5M calls via Call Metrics with average call duration of 6 minutes.

TechCrunch has actually written about both products – two examples are below:

“Google AdWords Advertisers Can Now Bid For Phone Calls In Search Ads”

“Google Mobile Search Ad Requests More Than Doubled In 2011”