Comments on: Dulance – We hardly knew you

By: Gabriel Hunter

Gabriel Hunter — Sun, 14 Mar 2010 10:55:28 +0000

Hi everybody!
I have a few things in mind about shopping comparison. I buy many products using shopping comparison websites and I think their are the best thing that happend to e-commerce; I mean let’s face it we all gone throught moments when we bought something and we didn’t get the exact product or we didn’t get it at all, althought we paid for it.

Shopping comparison essentialy help you to filter the stores to the ones that are serious and legal and to filter the prices, to buy the product at the best price.

There are many things to say about shopping comparison, but those are the basics (at least what i look for..)

Best Regards!

By: Bambarbia

Bambarbia — Mon, 16 Jul 2007 22:42:49 +0000

Hi,

Interesting article, and comments, and I have some thoughts… Tokenizer is crawler-based (sorry for copywriting); try to search AMD (or, for instance, Thinkpad X60) at Tokenizer and other similar sites… Only(!) 120 large stores are partially(!) indexed, and almost the same amount of search results comparing to competitors (including crawler-based)… I can add more stores, but need to tune engine at first, and to make website brilliant.

I’ll easily add category next week; it will be “Faceted Browsing” similar to shopping.com; it’s not a problem.

About product names: of course it IS possible to employ similarity algorithm , but if merchant wants to publish a suffix ***Free Shipping*** or even ***Refurbished*** I need to scrap it.

About “comparison”: currently, cheapest Lenovo prices are directly at Lenovo! (I mean top-notch products only).

Thanks

P.S.
this is second POST attempt, some SQL errors…

By: Todd Wilson

Todd Wilson — Fri, 31 Mar 2006 22:41:51 +0000

Sounds like you’ve already found a lot of success, Alex. That’s fabulous.

Handling comparisons definitely is one of the trickier issues we’ve run up against. Fortunately, we have a guy on staff who specializes in this kind of thing, so it’s just a matter of time. In answer to your question, yes, we’ll be allowing for comparisons even if it’s not an exact match. We’re anticipating it will be a type of spectrum of similarity. We could allow the user to set it, but we may also just display products that are 90% similar, for example.

Feel free to drop me a line if you’d like to chat about the screen-scraping aspect of your operation. If costs are an issue, we also offer a completely free edition of our software that may do the trick for you. My email address is my first name at screen-scraper.com.

By: Alex Firmani

Alex Firmani — Fri, 31 Mar 2006 14:14:28 +0000

The product name matching between retailers is a huge difficulty. My only method right now is manual– I have a script that does some intelligent matching, suggests the closest alternatives, and makes it a 1-click process. It is still quite tedious. I do think it depends on the niche though. I’m confident that if I were dealing in electronics or something else where all retailers use the same model/manufacturer numbers, the product ID matching could be 95% automated. With scrapbooking, I understand your dilemma. The brands might not be the same, yet the product is nearly identical– do you allow a comparison even though not an exact match? Tricky stuff.

The scraping is also custom written for each site. I’d never heard of a tool such as your screen-scraper and I will certainly look into it. As I’m expanding to other niches, I see myself working on something more advanced like that where there is less manual intervention for new retailers. Supposedly Healthpricer.com does a similar sort of scrape, taking in unformatted nutritional information and such. I am a one-man team though so my attention to each detail is limited. That does increase the profitability though– at only 8 months old, my site is more successful than Healthpricer.com with a team of 16 people. I’ve found a much better ROI by concentrating on high margin goods rather than feeling I must have every single product available.

By: Todd Wilson

Todd Wilson — Fri, 31 Mar 2006 05:13:57 +0000

Thanks for the reply, Alex. We’re actually in the process of enhancing the site to allow for comparison of identical or similar products. The difficulty is that, oftentimes, the same two products are named differently. As such, we’re formulating some complex algorithms for determining similarity between products by examining various attributes (including the name, of course).

It looks like you’ve done a good job on your site of allowing for comparisons. Do you mind sharing a bit how you went about it? How does your item ID system work? Are you manually associating products together, or do you have some way of automating that? Unfortunately, for us, automation is the only answer given that we currently have about 60K products in our database. Also, just out of curiosity, how are you handling the scraping? Have you simply written custom code for each site?

John, we’ve done a number of meta-search engines. You can check out our technology at http://www.screen-scraper.com/. If you’re interested in furthering the discussion, feel free to drop me an email. My address is my first name at screen-scraper.com. We’re probably in a good position to help you out.

Kind regards,

Todd Wilson

By: John

John — Thu, 30 Mar 2006 23:35:11 +0000

Hi,

I remember looking at Dulance.com a long time ago, but can’t recall their key differentiation points. Can anyone tell me what they were doing that was innovatiive? Also, I’m currently building my own comparison site, anyone know where a good price comparison script can be purchased? …or any experienced programmers interested in helping us build one?

Thanks,
John

By: Alex Firmani

Alex Firmani — Thu, 30 Mar 2006 15:49:49 +0000

Interesting model, Todd. Like you, I find niche shopping to be more rewarding than a general mall. I admire the ease of upkeep of such an automated data-entry system. It does leave me wanting a bit more as a consumer though– where are the price comparisons between retailers? It seems less of a comparison shopping service and more of a Froogle type data aggregator.

With my model (see http://www.supplementjudge.com/supplements/MuscleTech/GAKIC.html) I use both scraping and datafeeds; however, I integrate each one with my own item ID system– allowing price comparisons and easy upkeep of price changes. For my major sources of data/products, I setup scraping algorithms for quick imports of new products. The monetization model is purely affiliate.

By: Paul

Paul — Thu, 30 Mar 2006 14:16:46 +0000

i’m sure you already saw this, but just to reaffirm, the WHOIS database says Dulance is INactive but regisatry status is active (expires Sept 3 06)

By: Paul

Paul — Thu, 30 Mar 2006 14:00:23 +0000

will you be doing a story on the shoplocal/Cairo suit and subsequent closing of Cairo.com as well?

By: Todd Wilson

Todd Wilson — Thu, 30 Mar 2006 04:41:08 +0000

An excellent posting. This is very much along the lines of what we’re doing with our ScrapbookFinds site (http://www.scrapbookfinds.com/). We handle no data feeds, but instead crawl sites to populate our database. This means the data is very accurate and fresh. Our screen-scraper technology (http://www.screen-scraper.com/) also allows us to set up sites to be crawled in a relatively short period of time.

I’d be interested to hear about other revenue models. We currently primarily make use of affiliate programs, and have considered adding a PPC model. We also recently added AdWords, but are debating about keeping them. Has anyone else seen success with other methods?