Neotel gears up for corporate launch

Only five years behind schedule, South Africa's official second fixed-line telecommunications provider Neotel will be launching phone and internet services for business customers from mid-March. Apparently Tata-managed company has already acquired more than a dozen large corporate customers.

They are currently putting together a cable/fibre network and a Wimax/CDMA infrastructure that has the ability to bypass Telkom's stranglehold on home phone lines. Neotel are promising to undercut Telkom's prices (how hard can it be to undercut the highest prices on the planet?) and to beat them on service (again, how hard can this be?).

IOL Technology Article

Telkom Takes Over Internet Service Provider Africa Online

What do you do when you already own the telecommunications market in your home country? You expand internationally. Telkom has apparently acquired Africa Online, an ISP that operates in ten African countries, including South Africa's neighbours Namibia, Swaziland and Zimbabwe. Telkom will pay R70 million to African Lakes, the UK-based company behind Africa Online.

The penetration of the internet is of course very low outside of South Africa, and there is actually real competition to deal with, so Telkom may find its management competence a little stretched. In South Africa, Telkom's battles with its "competitors" could be likened to a club rugby team who only ever plays against school under-14 teams -- and gets to change the rules depending on the run of play. But out in the real world, there are All Blacks and Lions and more or less level playing fields. With luck, the experience will enlighten senior management at Telkom, and teach them something about customer service.

Perhaps they might even pick up on the economics of abundance -- imagine the impact on Telkom revenues of providing really low-cost web access in South Africa. Imagine the impact on the economy and education...

Business Day article

Nine Questions To Ask About Your E-mail Marketing Campaigns

The more e-mail marketing you do, the easier it is to slip into bad habits or to assume that what you know worked last year will work again this year. Here are nine questions you should ask yourself about your e-marketing campaigns, to try to keep them focused and effective.

1) Is your address list full of black holes?

Your opening rate may be low because you are sending to people who no longer respond. They are not bad addresses, just people who ignore you. Use your metrics to isolate the “black hole” addresses, those who have not opened your mail in more than six months. Then send them a targeted message with a compelling subject line, inviting them to update their mailing preferences, complete a survey, or even get some incentive. Then delete anyone who stays unresponsive.

2) Do you know how your messages look to all of your readers?

What impact will Microsoft’s decision to use Word to render HTML e-mail in Outlook 2007 have on the appearance of your work? How do your e-mails look on a mobile phone, web-based mail, Blackberry, or PDA? How do they come across in different browsers and mail clients? You should test in as many combinations of browser and mail client as you can. There are services that will do this for you. Try SiteVista's email service ($49 per month): you send an HTML email to the address they provide, and immediately view the rendered message as it would appear in various email clients. (To see how your message, or your site, looks on different browsers, use BrowserCam. A one-day pass gives you unlimited screen captures of your web pages using hundreds of combinations of browsers, platforms, resolutions and plug-ins – all for under $20.)

3) Do your opt-in processes actually work, and how porous are they?

The fewer steps a user has to take to get to the final opt-in click, the less likely you are to have people abandon the process mid-way. Test the links and every step that a reader will have to take. If there are five or more clicks from start to finish, you might have a problem. If you can reduce the steps that it takes, do so. Don’t try to get lots of information at the opt-in stage, merely an e-mail address will do – and you can get that automatically with a simple “click-here-to-subscribe”. If you need to know more about your subscriber, fill in the blanks progressively as your relationship matures.


4) Do all the links in your e-mail messages actually work, especially your unsubscribe links?

There are different types of links: internal links to content that allow readers to jump around the e-mail; links to content on websites; links to images not embedded in the e-mail; and service links such as subscribe, forward to a friend, and the all-important unsubscribe. Have someone test all of them before the e-mail goes out, and make sure the actual e-mail as received is tested on a computer other than the one used to create the e-mail.

5) Who is watching the incoming mail, especially in the unexpected mailboxes?

Just because you give people the right links to click dos not mean they will use them. Just because you tell people not to reply to auto-emailed messages does not mean they will obey your instructions. There should never be such a thing as an un-manned e-mail address.

6) How likely is your e-mail to trigger spam filters?

Different spam filters get alarmed by different things. Your images, links, hidden code, and even use of words or capitalisation can get your message bounced before it gets to its destination. Most e-mail marketing services provide an automated spam check of your mailing – use it. Services such as Constant Contact or Mail Chimp will alert you to things you ought to change to increase the likelihood of your e-mail getting through.

7) How does your e-mail display in mail clients using image-blocking and/or preview panes?

Most mail clients (even web-based mail) now provide preview panes that allow users to see the top part of a message without actually opening the whole thing. That’s good because you are not solely dependent on your subject line to get the recipient to open the message rather than delete it; it’s bad because the top part of your e-mail may be a turn-off. At the same time, most mail clients’ default security settings block images in e-mail (whether embedded or linked). So it’s a good idea to use a combination of compelling content and non-image layouts, colours, and fonts to make the first five-ten centimetres of your message as irresistible as possible. Once the reader wants to see your message, they will do the simple click-to-enable the images. To help them decide to do this, make sure you use compelling alt tags with each graphic which will display in place of the image.

8) Are you immediately engaging new subscribers to your list, and trying to ensure those who have recently unsubscribed leave with the best impression of you?

Don’t leave a new subscriber hanging for months after they have made the commitment to join your list. OK, some people fall through the cracks (we all do it), but work on making your cracks a lot smaller. When someone unsubscribes, confirm this in an e-mail or on the unsub landing page, and provide a link to a brief exit questionnaire. You can learn a lot from this feedback that can help you recapture those who you have lost, or prevent more people from leaving for similar reasons. Remember, markets are conversations and everything we do should be geared to facilitating relationships.

9) Are you sending messages too often, or not often enough? Are you sending them at the right time of day, on the right days?

Depending on what it is you are communicating, you might be mailing daily, weekly, bi-weekly, or even seasonally. If your mailing goes out every few months, you risk people not remembering you and unsubscribing. Too frequently, and they will unsubscribe because they feel spammed. The key is to match content to frequency, and frequency to recipient expectations. If you have a large enough list, test different frequencies to see what the success rate is and what the unsubscribe rate is.

Personalization Desire Outweighs Security Concerns

It is hard to imagine that people who are neurotic about data loss and identity theft would willingly relinquish security for convenience. But time and time again we see consumers acting in a way that might make life harder tomorrow in order to make life easier today. In German supermarkets, people who would refuse to allow their government to capture their fingerprints willingly pay for groceries by pressing their finger on a fingerprint scanner at the till.

The recently publiched "2006 ChoiceStream Personalization Survey" found that over half of the respondents will provide demographic and other personal information in exchange for a personalized online experience.

Part of this is familiarity. Ten years ago when online sites like Amazon.com first started using cookies and subscriber data to customise the shopping experience, people were wary. Suspicions abounded about who might get hold of the information and how it might be abused. But if you don't provide the necessary information, and if you crush your cookies regularly, the online experience can be dry, tedious, and often irrelevant to your interests. Over time, people get accustomed to the personal service and become more willing to give away information. At the same time, they get more sophisticated and discerning, and they actually read privacy policies and make conscious judgement calls about who to trust and who to avoid. The more information consumers are exposed to online and the more decisions they are making online, the more dependent they become on personalisation as a data filter.

According to the report, in the past year, consumer interest in a more personalized experience has increased by 24 percent to 57 percent of respondents. Consumers willing to let a site track clicks, purchases, and other behavior increased by 34 percent in the same period. Concerns over personal data security remained high, with six out of ten people expressing concern.

Not surprisingly, younger age groups have less concern about privacy and more interest in personalisation. Comfort with the norms of social networking sites play a role, with younger people expecting sites to provide personalised recommendations.

original ClickZ article

Advertising Boosts Municipal Wi-Fi Ubiquity

There's nothing new about municipal WiFi. It has been around in the US for at least five years (that's 28 internet years), and longer in Japan and South Korea. Now it's available, or in process, in more than 300 cities across the US, and it is becoming almost the norm in major European cities.

The basic idea is that a municipality decides that broadband internet access is a utility, much like electricity or schooling, and that making it available to everyone within the city limits for free can only be good for the city's economy and education levels. The democratisation of information is a strong political driver, and a pretty powerful economic imperative. So the municipalities put up wide area WiFi networks that anyone in town can access with a WiFi enabled computer, phone, or PDA.

Typically these services are free, or at least they are paid for out of taxes, or they are advertising supported. And they are not that expensive to set up: Philadelphia budgeted only $10 million to WiFi enable the whole city, with ongoing costs of about $1 million a year. That's a small price to pay compared with the potential benefit to small businesses, learners, and government.

In South Africa, Knysna has muni WiFi, and there is talk that Gauteng may WiFi-enable the greater Johannesburg area in conjunction with iBurst. There are, of course, political considerations and pressure from internet providers who see their lucrative markets under threat. But inexpensive muni WiFi is inevitable even in South Africa.

Often muni WiFi is paid for by advertising: when you log on to the service, you see an ad. Often these ads are very local, for businesses within range of the nearest signal source. That produces great opportunities for advertising services for small businesses that otherwise would never consider marketing online: local restaurants, dry cleaners, finacial advisors, or car dealers can communicate with potential customers who are right there in the neighbourhood. Try that with TV or local radio.

In the US, a hotspot and advertising company called JiWire now plans to offer ad-supported Wi-Fi through a relationship with Ultramercial. The ads will appear prior to gaining free Internet access at hotspots. To avoid the initial ads you can simply pay a small fee to get online. Once WiFi providers and income generators like JiWire start investing in muni WiFi infrastructures instead of waiting for municipalities to take the initiative, things start to develop at a very rapid pace.

There's an article on JiWire here

New Google AdSense Policies

AdSense is the Google service that allows websites and (more typically) blogs to host small pay-per-click text advertisements that are dynamically served to their pages depending on the context of the page's content. Keywords that happen to appear on the page trigger the specific ad: write an article about boats, and an advertisement for boats will appear in the Google AdSense placeholder. If a reader clicks the ad, the page owner gets paid a piece of Google's pay-per-click fee.

As with everything Google, there are policies. The AdSense policies have now been updated and made more reader-friendly, having not changed since March last year. Here's what's new:

By far the most significant change is that Google no longer allows its AdSense ads to appear on any site (not just page) which also hosts other non-Google advertising that is formatted to look like AdSense advertisements. Here's their language: "In order to prevent user confusion, we do not permit Google ads or search boxes to be published on websites that also contain other ads or services formatted to use the same layout and colors as the Google ads or search boxes on that site. Although you may sell ads directly on your site, it is your responsibility to ensure these ads cannot be confused with Google ads."

Effectively, Google is enforcing a design trademark on its ads, presumably because it feels that the "look" of its ads conveys some kind of quality to potential clickers, or more likely because it wants to maximise the click rate by eliminating competing ads,particularly other contextually targeted advertisements.

There are other changes that affect sites who host AdSense ads. Sites may now place up to two AdSense referral links or buttons on a page per product referred. And referral offerings must be made without requiring users to provide any information such as e-mail addresses from users.

Many blog authors make significant amounts of money from AdSense, and will go to extraordinary lengths to increase the click rate and therefore their income. So Google is now also regulating the use of images next to AdSense ads. So sites may not draw attention to ads by, for example placing large "Click here" arrows next to the ads, or place any images next to the ads designed to mislead users.

If you use a Google search box on your site, you can now also place an AdSense ad placeholder on the search results page.

Finally, Google has broadened its previous policy which forbad the placement of AdSense ads on pages that contained any media such as MP3s or newsgroup postings to include "web pages with content protected by copyright law unless they have the necessary legal rights to display that content." So don't expect to see AdSense ads on YouTube any time soon!

Google Updates Its Ad Quality Scoring

Google is rolling out an improved ad ranking algorithm for AdWords, with a little more transparency than before. They actually now give you a vague idea of what your ad quality score is.

AdWords is the system that places small text advertisements on a Google search results page or places text/image ads on pages in thousands of AdSense enabled sites around the world. (This is relevant for South African advertisers, because Google lets you target the pages on which your ad may appear by location -- so your ad could show exclusively on South African websites if that was your wish).

Those ads are designed and created by advertisers, and are associated with keywords that the advertiser "bids" for. When a searcher enters a search phrase that is or contains a keyword that an advertiser has bid for, the relevant ad appears as a "sposored link" in the upper right hand corner of the search results page.

But many people will bid for the same keywords, so several ads may appear. And the ad at the top of the list is not necessarily the one whose creator has bid the most for the keyword in question.

The Google algorithm takes more than just bid level into account when deciding which ads get precedence -- the "quality" of the ad plays into the ad ranking as well. (If that sounds confusing, it's because you are not supposed to know how it really works -- that way advertisers can't cheat the system and Google's competitors can't easily copy it).

One of the elements that go into the AdRank calculation which scores the quality of your ad is the quality of the landing page -- the page searchers get to if they click your ad. (The other elements are the actual click through rate your ad gets on Google, and the relevance of your keyword and ad text). In theory even if your bid is low you can have your ad displayed at the top of the list if you have a very high quality score, say because you have a good landing page. And, if your quality score is high, you can find yourself paying less per click than the ads placed below you.

This is because Google does not charge you the maximum you have bid for a keyword. You pay only one US cent more than you would have had to to pay if you were the next highest ranking ad.

Here's an example of how it works. Remember that you never actually get to know what your quality score is -- Google will tell you that it is good, average or bad but will not disclose the number. In any event, the score is recalculated every time a relevant keyword search is run.

Joe's Garage bids a maximum cost per click of $0.40 for the keyword "fanbelt"
Jill's Garage bids $0.65 for the same keyword.
Pete's Garage bids only $0.25.

Joe has a well worded ad that gets better click rates and has a good relevant landing page. Jill has not taken much care over her ad, doesn't get a lot of clicks, and sends clickers to her home page. Pete is somewhere in between. Google does its quality calculations and produces Quality Scores of 1.8, 1.0, and 1.5 for Joe, Jill, and Pete respectively.

Google multiplies Joe's maximum bid by his Quality Score and gives him a ranking number of 0.4 x 1.8 = 0.72.
Jill's ranking number is 0.65, her maximum bid (0.65) times her Quality Score (1.0).
Pete comes out as 0.25 x 1.5 = 0.38.

Joe goes to the top of the list, since he has the highest ranking number, with Jill second and Pete third. But what do they actually pay each time their respective ads get clicked?

For Joe to be listed in second place he would have to pay Jill's ranking number divided by his quality score (0.65/1.8), or $0.36. But he's ranked above her, so he must pay one cent more, or $0.37. Note this is lower than the $0.40 he actually bid.

For Jill to be listed in third place she would have to pay Pete's ranking number divided by her own quality score (0.38/1.0), or $0.38. But she's ranked above Pete so she pays a cent more, or $0.39. That's a lot less than she bid, but it is more than Joe is paying in the top slot.

Pete has nobody below him, so he pays only his minimum threshold fee (in this case only 4 cents) every time someone clicks his ad.

The moral of the story is that you can really lower your CPC rates by making sure you create good ads and use focused relevant landing pages. And being third on the list can be a really cheap option..

To become an instant expert on AdWords, go to Google's AdWords Learning Center".