Loki – interesting, but can it work?

29 May

So I came across a post on Loki over at Screenwerk, and decided to take a closer look.

The premise is interesting – using wifi triangulation it is able to discern where a user is located. This is great for anything localized – eg local search. From our own experience a lot of users don’t even bother with entering in the ‘where’ field – what they expect to get, I still haven’t figured out.

So – you have to download the toolbar for it to work (thankfully with both Internet Explorer and FF versions). The toolbar takes a while to install, and for a desktop computer, it falls back to figuring out where you are based on IP. It would have made sense that, once realizing this was not a wireless computer, it would have prompted me for my address. Instead I had to initiate this process. Even more – I am very concerned about my privacy. I would have appreciated it if I could have just entered my ZIP code – instead it demends a street address.

The system is a bit sluggish – adding channel seemed to freeze up my computer, as did searching anything through the toolbar.

It works well. I went ahead and submitted iBegin Source to be included in their list of search engines.

But the real issue here is their new ‘Javascript API’. Using this API you can do ping the loki servers to figure out where a user is (provided he/she is using Loki). With the latitude/longitude of the visitor, you can easily serve up localized content.

But – it seems very muddled. The code itself is really simple – a few lines where you connect with the Loki server and extract this information.

But there are too many negatives. First of all – if Loki.com goes down, your site performance will degrade (due to the JavaScript call). I can trust Google will be up (for Google Maps) – but Loki?

Second of all – the restrictions seem rather arbitrary. We are given 10,000 location transactions per day – what is a transaction? Is one transaction one lookup (even if the user isn’t using Loki). Or is it only a transaction when the user is using Loki?

Lastly – the badges page. Provided we can get over the previous two issues, if I want a user to utilize Loki, I don’t want to send them to a common page. They are my visitor – I don’t want to be sending them to a competitor’s site. For the badges to work, they must provide a ‘clean’ landing page with minimal distractions. A link at the end saying ‘Visit Loki.com for more location-enabled sites’? Fine. More than that is too much.

At the end of the day, what I am curious about is – how many people are actually using Loki? Getting iBegin Source to work with the toolbar took all of 2 minutes – just do a check if latitude/longitude are set. But utilizing the JavaScript API requires a lot more work – are there enough Loki users to make it even worthwhile? While I hate to use those online ranking sites, the early indications are not so good.

2 Responses to Loki – interesting, but can it work?



May 31st, 2007 at 5:32 pm


I just read your post of the new Loki Javascript API and thought I would respond personally. I am in charge of Loki as a product and I take your comments very seriously.

First of all, you make very good points. We launched Loki almost exactly a year ago as the first toolbar to let users start using location-based services on their laptops without needing a GPS device. Over that year we have tried to improve the product, streamline the features that we push out there and learn more about our end users.

The Javascript API is a feature that we hoped would help take our location technology and put it in the hands of web developers so they could start to do something meaningful with the information and begin to deliver more enhanced services around location. So while its a great first step, we definitely need to keep working at improving the product to better serve the people who want to use the API on their site. This includes bringing the download size down and not pulling users away from the site they were currently on just so they can location-enable that partner site. We want to support our partners in the best way possible, so your feedback is great to hear. So please stay tuned for the next release as you should see a lot of these issues are being resolved.

As for the 10,000 transaction restriction that was put in place to try and find a happy medium between non-commercial users and commercial users. If you look at the Yahoo! Geocoder API, they limit you to 5,000 requests per day — we thought we were being good samaritans by doubling that number. Is there a number that you have in mind that makes more sense? We are definitely willing to listen.

So know that we are committed to improving the product and we need feedback like yours to do it. So I really appreciate the post and I would be happy for you to email me so we can discuss it more.


Ryan Sarver



May 31st, 2007 at 6:24 pm

Thanks for dropping by Ryan.

One of my recent posts is about how companies that embrace all the popular platforms abound will be the ones that will win. So I am very intrigued by what Loki can do – removing the ‘where’ field in local search is quite big. It would be great if you could notify me when the new release does come.

The 10,000 restriction – there is a reason why we built our own geocoder :) I’m not saying 10,000 as a number itself is low – it was the wording that caught me. Perhaps I mis-interpreted, but it sounded like if you went over 10,000, you were (essentially) SOL. I understand you don’t want to be hammered, but language that conveyed your willingness to work with a partner like me would go a long ways towards making me feel better. After all, it is a win-win situation here – the more users I have using Loki – well, the win here for you guys is obvious :) [and you didn't clear up what a transaction was - any call, or only successful calls].

All in all – I am intrigued. Honestly the biggest thing holding me back is the actual adoption rate of Loki – any chance sharing some numbers on that?