Minor bugfixes 1st November 2018

After fixing the problem with our SSL certificate we took the opportunity to look into a few more bugs that have been present for a while but never properly addressed, mostly to do with the PriceTracker:

  • PriceTracker storing 0.00 prices:Ā for whatever reason, some vendors report prices of 0.00. Obviously they are not giving anything away, there is probably some other bug at their end, but there is not much we can do about this except exclude such products both from the price comparison listing and the PriceTracker database. The latter was not being done properly, so $0.00 prices were still getting in there. We think we have solved this problem now. Another problem that MIGHT need to be looked at in the future is when a unit price is very low, say $0.001, then it may be displayed as $0.00 and excluded. Such prices are supposed to be rounded up, but there is a possibility that we have missed some cases where they are not, we will keep a lookout for those.
  • PriceTracker displaying a null price for new products:Ā products for which price data was being retrieved for the first time were displaying a nullĀ Lowest recorded price: – annoying, but could never quite be bothered to fix it. It’s done now!
  • Other behind-the-scenes stuff:Ā of no interest to users, but we changed the way we link to Amazon in order to stay in their good graces. Amazon links are now no longer redirected through Skimlinks but through our own ‘URL prettifier’ , basically so we can continue accessing Amazon data and earning commissions from any referral sales.

Other issues to be addressed

The PriceTracker relies on data supplied by retailers, so we cannot actually know whether the price data is accurate or not – we often get weird fluctuations in price, like an item that was $100 reduced to $10. Such a scenario is unlikely in ‘real life’ – but not impossible, so what do we do? Not much really, we could trim any outlier prices but what if they are a massive bargain! You never know! So we just leave them in there for now.

A reverse example of this is when the price of a product is unfeasibly high. Check out this cute chappie here. Apparently he was selling for $129,999.00 on 2 October 2018! Does seem a little on the high side! Of course that is a typo or something on the part of the seller, but how do we tell what the ‘real’ price was – the high one or the low one? Software isn’t clever enough to determine that – a historical average would give us the probable range of prices, but how to determine where the cut-off is for outliers, and what to do when we do not have enough historical data? For now we just leave these in there, it’s kind of fun anyway!

Leave a Reply

Your email address will not be published. Required fields are marked *