The sponsored lunch today at Search Insider Summit comes from Microsoft adCenter which included a demo of their new Keyword Services Platform and add-in to Excel 2007.

Stacy Harris:

Feedback from marketers is that traffic is good, but they want more of it. Marketers also want AdCenter to be easier to use.

Natala Menezes is up next and promptly announces Stacy was engaged last night!

Key points for current and future development of adCenter: Quality, Transparency and Simplicity.

What does this mean to keyword research, optimization and analysis? Access to real data, not just graphs including demographic and geographic information. It also means easy to use tools including keyword libraries and an adCenter add-in for Excel 2007 which focuses Keyword research, forecasting and monetization data.

Next up is Zhaohui Tang.

Keyword research is the most important aspect of any search optimization or analytics effort affecting: expansion, monetization, commercial intention, seasonality, synonyms, categories, translation, demographics and others.

Microsoft adCenter Keyword Services Platform: Algorithms to extrapolate new terms from a set and categorical relationships. Server platform offering keyword related services for adCenter but also for 3rd party developers to create new advertising applications.

Demo of KSP. I was sitting at a very bad angle as you can see, so I didn’t get any good photos. While the tool will not be released until early 2008, hopefully the adCenter team can give me access soon to do my own demo and review. Hint hint.

The demo shows Excel 2007 with numerous tabs on top showing a variety of keyword applications that go to the web and bring back the data directly into excel. Keyword extraction, suggestion, search buzz, monthly traffic, keyword categorization, geographic info, demographic info, monetization info (clicks, impressions, avg positions, CTR, CPC). There are also advanced algorithm settings.

The app is cool enough that there are whistles from the audience after the search volume tool is run resulting in big smiles from Stacy and Natala:

Audience: Will there be backward compatibility for Excel 2003?

Natala: Microsoft is working on a web version and the MSFT engineer fellow said there will be a backward compatible version for Excel 2003

Audience: (Gord Hotchkiss) are all the features you showed only for your “head” phrases? How many terms will be returned?

Engineer: Only keywords where at least 10 queries per month. There is also a “generous” quota of query volume.

Natala: It’s also a live database being added to and will grow as time goes on.

Audience: How is the confidence score determined?

Engineer: Long answer, but basically algorithms that give an indication of probability.

Comments

Session description: Expanding and optimizing keyword lists doesn’t need to be complex or time consuming. This session will provide a preview adCenter’s latest innovation – the adCenter Add-in for Excel 2007 and how it can be used to rapidly and easily build out or expand keyword lists and plan keyword strategy.

Presenters:

• Stacey Harris, Agency Marketing Manager, Microsoft
• Natala Menezes, Product Manager adCenter
• Zhaohui Tang, Senior Group Program Manager

Natala Menezes
The goals and direction for adCenter:

• Quality
• Transparency—real, actionable, accurate data
• Simplicity

When we talk about transparency, we want to give you access to real data—not just graphs and relative measures. We want them to be actionable and performable for you. We’re big on demographic and geographic info.

Simplicity—easy-to-use tools in the environment you’re working in already.
New product launching next month inside an application that you’re already familiar with, Excel: the Keyword Service Platform (KSP).

What is it? A treasure trove of keyword info. It has algorithms to extrapolate new terms from a set, define categorical relationships, access to real data. It’s a development platform.
(Want access to SDK? adcafb@microsoft.com)

“Better Together” [because everything at Microsoft needs a name.]

• Search marketers already work in Excel.
• Keyword research shouldn’t involve so much cut & paste.
• Excel 2007 advances: unlimited rows; advanced add-in platform
• You don’t need to be an expert in Excel to use these tools
• Data is tuned for adCenter but can be applied to any campaign on any engine

The right keyword for the right audience to capture the right customer
3 philosophies: keyword research (easily generate and expand your list), keyword forecasting (historical monthly data and future predictions), keyword monetization (understand KPIs to help you decide which keyword to use & how much to bid)—CPC and CTR

Zhaohui Tang

Keyword technologies & related tools

Keyword research is the most important aspect of any search optimization, promotion campaign and web analytics.

expansion—relationships

monetization

commercial intention: map vs. digital camera. We’ve had 2 researchers working for 2 years on this.

Seasonality

synonymy

categorization

translation—automatic
demographics, geographic

KSP is a server platform that provides a set of smart keyword-related web services to empower adCenter, windows Live and 3rd party developers building next gen of online ad apps and beyond.

Demo and XBOX give away

XL plugin—ad intelligence. Lovely icons: keyword wizard, keyword extraction, keyword suggestion (options), search buzz, monthly and daily traffic, keyword categorization, geographic, demographic, monetization and vertical kpis, advanced algorithm.

Daily data updates, researchers in adCenter labs building new models—all at your fingertips

type in URL (autos.msn.com)—generates keywords (as many as you want). Gives you top two most likely categories (segments): vehicles/automobiles. Financial services/insurance/vehicle insurance
select keyword, use expansion. Finds all the keywords containing those keywords.

Easy to delete keywords

100s of keyword in seconds—confidence for relevance (r2)

Monthly traffic and forecasts—real numbers!

Daily traffic and forecasts (great for dayparting)—as recently as 2 days ago.

(Conservative approach—take out fraud/bots)

Geographic info—where are people searching from. City, state, country

Demographic info: percentages male/female, age distribution (over a set date range) (some uncertainty, of course—margin of error)

Vertical KPIs—compare insurance vertical v. used car vertical—clicks, impressions, average position (I guess your ads’), CTR, CPC

Monetization—show CTR, CPCs, etc.

Algorithm settings: generates a chart with different positions over time—and can graph impressions, CTR, CPC, etc.

Conclusion: Keyword Wizard is fabulous
available in early January
acpc@microsoft.com to get access to tool

Other features to play with:

search buzz—reveals keyword trends, “spiky” occurrences. Great for capturing popularity data, top frequent keyword over a longer time period

Questions

Is there going to be back end compatibility for non2007 XL?

We’re debating whether to provide 2003 version or do web-based. It would be pretty significant to do 2003. (Zhaohui says they’re working on it, but have no dates)

If I’m not an XL user, can I use KSP? Do you buy/subscribe for access?

We’re working on this as FREE tools (except for Excel).

All the features we saw—are those available for all keyword? Or are there thresholds?

Zhaohui: There are certain things (privacy issues). We only do keywords with at least 5 customers bidding on it. There’s a quota system to avoid abuse (but it’s fairly generous). For larger deals, you can get partnership.

To clarify: is there search volume info threshold?

If there are at least 10 queries a month, we’ll show then numbers.

It’s a % based system on Passport data. Some confidentiality on geographic and demographic info. It’s a confidence score to use as a predictive measure.

Can you explain how the confidence measure is performed/conducted?

Zhaohui: Each algorithm owner decides how to interpret the confidence score—depends on the researcher that wrote it.

Natala: More likely to vary in predicting CPC. You can use this for establishing a testing pattern.

They’re working on lag times in budget/updating.

The search volume data is just from Live. Right, so if you wanted to extrapolate that, you should use market share data. . . .

Natala: We often say the data is best tuned to adCenter, but you can use it anywhere.

Comments

I wonder what Rick Segal, for instance, would say?

Update: Rick just blogged his answer.

Joi Ito? Allen Morgan? Bill Gurley? Jason Caplain? Steve Hall? Steve Jurvetson? Jeff Nolan? Tim Oren? Ed Sim? Martin Tobias? Fred Wilson? Brad Feld?

Now, I’ll be honest, most of them will say Microsoft is dead. That open source is where the action is.

Why do I know that? Because they’ve told me such. And, because most of them are looking at yesterday’s money and assuming that tomorrow’s money is gonna be made the same way. Google made some of them fantastic sums of money, but now the rest are chasing that dream.

Problem is, will tomorrow match up with yesterday?

You’re probably noticing a theme here. I’m posting all of the bad news about Microsoft. Steve Rubel is switching to a Mac. The VC firms are supporting open source. Longhorn is seen by many as a failure waiting to happen.

I love big challenges. These are big challenges.

But, I keep going back to what’s important: the small things. The details. The devil is in them, right?

Obviously I’m speaking from a position of strength here. I know what’s coming. So, I’m fearless. Go ahead. Punch us in the stomach. Again and again. Go ahead. Tell us we’re out of the game. Go ahead. Tell us we can’t do anything right.

Let’s get it all out, all here, all now.

Give us your best shot!

LInk: Is .Net Failing to Draw Venture Capital Loyalty?

Reader Comments…

Robert Scoble is the founder of the Scobleizer blog. He works as PodTech.net’s Vice President of Media Development.

Go to Scobleizer …

SIM Access Profile basics

The SIM Access Profile functions in two modes, client and server, and can work in two types of scenarios. Both of these scenarios involve Bluetooth-enabled server and client devices that together have an established Bluetooth connection. In the first scenario, only the server has a SIM card and the client establishes a connection with the server. In the second scenario, the client has a proactive SIM card and the server has the additional SIM card; both the server and client initiate a connection. Figure 1 shows a general description of the Profile’s workings.

Figure 1. The SIM Access Profile using server and client mechanisms

Figure 2 illustrates a standard SIM Access Profile configuration with respect to the Bluetooth stack.

Figure 2. The SIM Access Profile using Bluetooth

The client can initiate the following operations to the server (with the SIM card):

• Manage connection. Establish and terminate a SIM Access Profile connection.
• Transfer APDUs (Application Protocol Data Units). The application sends command APDUs, while the SIM sends response APDUs. Command and response APDUs only occur as pairs, in which each command APDU is followed by a response APDU.
• Transfer ATRs (Answer to Reset). Sends the ATR content (which says that the receiving side has accepted the APDU) from the server to the client over the Bluetooth link.
  Control the SIM. Turn the SIM card on or off.
• Control the SIM. Turn the SIM card on or off.

The server initiates the following operations to the client:

• Report status. Server informs the client about their connection status.
• Transfer card reader status. Sends the card reader status to the client over the Bluetooth link.

Security
SIG proposes four levels of system security: bonding, baseband-level encryption, server-initiated authentication, and link keys. You can find more details about these levels in the SIM Access Profile specification (see the Resources section later in this article for a link).

The intelligent telephone booth scenario

Imagine a scenario in which a telephone booth has a Bluetooth-enabled SIM card and a user has a mobile device unconnected to the cellular network, but running a SIM Profile client.

Upon approaching a mobile device cell booth, the user would simply establish a connection, using his mobile device, and make a regular call. (The cost for airtime and local call charges for using a cellular phone is much higher than the cost of a call made using such intelligent booths.) The amount would then automatically deduct from the user’s SIM card by transferring APDUs that modify the card’s ACM (Accumulated Call Meter, which accounts for the amount transacted during each call). Figure 3 illustrates this process.

Figure 3. A telephone booth that uses the SIM Access Profile

There are a couple of reasons for using the intelligent telephone booth:

• Each telephone booth could cater to multiple users provided there are multiple connections, thereby saving a lot of space.
• As of today there is a huge disparity between the land line costs and cell phone costs. This mechanism combines the cost of using a land line, providing the user a flexibility of using Wireless communication within a particular radius from the booth.

The intelligent vending machine scenario

To enable a mobile customer to purchase a vending machine item with her cell phone or PDA, the device would need to:

• Contain a SIM card and subscribe to a telephone network
• Be Bluetooth-enabled
• Support the Personal Area Network (PAN) or LAN Access Profile
• Support the SIM Access Profile

The PAN Profile establishes a simple connection between the vending machine and the user’s mobile device. An application runs over the PAN Profile and asks the cell phone vendor for various user authentication inputs. For example, it could ask what type of soft drink the user wants, and then how many of that type.

The payment gateway machine, namely the vending machine, then establishes a SIM Access connection (as a client) with the SIM Access Profile’s server residing on the mobile device.

Authentication is provided if the payment gateway connects to a cellular network in the same way described in the GSM 11.11 Specification, Section 7.1 (see Resources). Figure 4 illustrates the architecture of such a system.

Figure 4. Architecture of the Bluetooth-enabled vending machine

Transaction using a Bluetooth module

Alternatively, you could configure the SIM Access Profile client to send the APDUs, which disable and enable the card holder verification (CHV), to the SIM Access Profile server. The Bluetooth module on the vending machine would act as that server.

First, you would send the Disable CHV (Card Holder Verification) APDU, described in Tables 1 and 2, to the server’s corresponding SIM card. This could require that the user enter a PIN into the application screen and then the application on the server would disable the CHV (in order to modify the ACM entity file, which is described later).

Table 1. Command for disabling CHV

Table 2. Description of the data field for disabling CHV

You could also use biometric identification mechanisms on mobile devices to disable the CHV. This allows the SIM Access client to access and modify various entity files:

• ACMmax Value. The Accumulated Call Meter Maximum Value could be the maximum value the SIM permits at any time; for example, in situations where pre-paid cards are used.
• ACM. The Accumulated Call Meter tracks the amount of money transacted for each call.
• SPN. The Service Provider Name entity file stores the name of the service provider. At regular intervals, the vending machine would also need to send this data to the back-end service provider to inform it of transactions performed.
• PPUCT. The Price Per Unit and Currency Table is useful as a currency conversion table to deduct the appropriate amount of money when traveling in another country.

A simple SIM Access Profile transaction would be to increase the amount of money from the ACM entity file; for example, to make the corresponding transaction equal to the number of calls of the same value. Say I order two soft drinks and each costs one dollar. If each call I make on my cell phone costs me 50 cents per minute, the APDU would deduct the amount equivalent to four calls for having ordered two beverages. Once the value is deducted from the SIM, it is displayed on the Mobile Equipment User interface using the PAN Profile. As soon as the user accepts the transaction, the transaction happens; for example, the vending machine dispenses two soft drinks.

Once the operation completes, the Enable CHV APDU command, shown in Table 3, is sent to the network. This enables the CHV and prevents any unauthorized modifications to the ACM.

Table 3. Commands to enable CHV

To reflect the billed amount, you could use one of the following mechanisms:

• Wireless transaction. Each transaction is immediately pushed from the vending machine to the back-end cellular network. The vending machine, however, must provide the cost for the airtime used to communicate with the back-end network.
• Wired transaction. Because the transaction happens over a wired network, you would need additional network infrastructure support. However, this could be a cheaper method.
• Delayed transaction. The vending machine or monetary service provider initially holds all the details in back-end storage and flushes the locally stored data at a later time, say at the end of the day. Because vending machines are of a limited capacity (each vending machine can hold only a maximum number of soft drink bottles), the software could provide sufficient memory to store all transaction details. This would eliminate the need for any new immediate transactions.

Summary of a typical SIM Access Profile transaction

Here are the steps you might take in enabling a user to use a mobile device (that has a SIM card) to pay for the cost of a vending machine item:

1. A user approaches the vending machine with his Bluetooth-enabled PDA.
2. The vending machine, which has a Bluetooth module and associated software, initializes a connection to the PDA’s SIM Access server using the PAN Profile. (The PDA is given a link local IP address so its packets can be sent to and from the vending machine.)
3. The user’s application screen requests for a password/biometric verification to enable transactions. The verification sends a CHV Disable to the SIM card.
4. The applications on the PDA and the Bluetooth-enabled vending machine start to communicate over PAN Profile; for example, the user selects a particular item.
5. The vending machine’s SIM client sends the corresponding APDU to access and increment the appropriate entity files to make the necessary deductions.
6. Upon receiving an acknowledgment APDU, the SIM Access Profile indicates to the application that the transaction is complete.
7. The vending machine dispenses the selected item.
8. The network records the transaction from the user’s SIM card by reading the entity file.
9. Finally, the vending machine sends confirmation of the transaction to the PDA’s Network Service Provider.

Note: The company that provides the Bluetooth-enabled machine can determine how to deploy the application that sits on the mobile device; for example, the application that sits on top of the SIM Access Profile and the PAN Profile could be downloaded onto the device from the Internet.

Conclusion
Combining the Bluetooth and GSM specifications enables users to perform monetary transactions using their Bluetooth-enabled devices. As a result, you reduce the time and cost of a transaction and eliminate the need for carrying cumbersome change. Because these transactions are secure, with multi-level security protection, they can help usher in electronic currency or the SIM. From use in ATMs to parking lot meters to automatic bill payment and beyond, this mechanism offers endless use possibilities as well as a glimpse into an exciting future.

Resources

• Find both the SIM Access Profile Specification and Bluetooth Standards Specification on Bluetooth.org.
• The GSM 11.11 Specification shows how to connect the payment gateway to a cellular network.
• Check out the IBM Linux Watch with Bluetooth capabilities.
• Also read Hemang Subramanian’s “Handing Off to Bluetooth” in the developerWorks Wireless technology zone. (developerWorks, December 2002)

First published by IBM developerWorks at http://www.ibm.com/developerWorks/.

Hemang C Subramanian holds a B.S. degree in computer engineering from the Karnataka Regional Engineering College, Surathkal (now known as the National Institute of Technology, Karnataka). Since 1999, he has been employed with IBM India, where he works on various networking and wireless networking technologies. He has one US patent pending and writes articles for various international publications and IBM developerWorks. He can be reached at shemang@in.ibm.com