5

Kijiji Canada is rolling out a new system for merchant accounts.

You create ads, and put a bidding price on them, much like adwords, but it's based only on the words in your title. (Possibly the first hundred words in thd ad too) No separate keywords.

The ads winning the auction will show up in the 4th, 11th and 17th positions on the page. You can a sample page here:

http://www.kijiji.ca/b-home-outdoor/edmonton/c19l1700203

I signed up, and created 3 ads twice, once as a private individual, and once as a merchant.

Merchant ads are distinguished by a pale lavender background, and a logo on the right hand side.

After one day the private ads received 12 clicks. The merchant ads received 224 clicks. I think that either Kijiji's software is broken, OR they are inflating the numbers.

I emailed my account rep asking for some third party verification. No Joy. The analytics they provide are all that is available.

Screen of results merchant account Screen of results private account

I have paused all ads in my account until this is resolved -- either an explanation as to why the merchant ads get such disproportionate attention -- or a method of 3rd party audits.

I have suggested that a reasonable audit method would be load one or more images off of my web server from a page labeled "no cache" If the image is fetched with a parameter, or alternately is fetched with a name foo+sequence#.jpg, and the server rewrite it to foo.jpg this would keep it from being cached.

What is the best way to see that you are actually getting the clicks you are paying for?

Edit: My smell-a-rat is going off full tilt:

Kijiji.ca according to HypeStat (http://kijiji.ca.hypestat.com/) gets about 10 million hits per day. Edmonton metro area is 1.1 million. But lots of nearby towns. Let's call it 1.5 million

Canada is 35 million. So edmonton is 4% of Canada. So 4% of the traffic should be for Edmonton's kijiji. Ok. I'm ignoring people who search Alberta.kijiji.ca. So we have 400,000 edmonton hits per day. If all page hits are ad pages (Ignoring people using the help pages, etc) and each page shows 20 ads, then there are 8 million classified ad impressions per day. It might be more than this. I've heard that for any given city either Craigslist or Kijiji dominates. So it might be double this.

There are 333,000 ads on edmonton.kijiji.ca 176,000 on buy and sell 3,800 on home-outdoor 193 on plants, fert. and soil.

Yet the claim on the screenshot above that my ads had 60,000 impressions. MY ads had 0.75% of all the impressions you made. But the category of ads has 193/330,000 = 0.058% There is a factor of 12 difference. Do you expect me to believe that there are 12 times as many people browsing this category as the average for the entire site? This is NOT the time of year that most people think about garden stuff. I would expect the ratio to be the other way.

All three of my ads were on page 1. So this means that page 1 of plants,fert,dirt was loaded 20,000 times in just over a 24 hour period. (This assumes browse, not search) That this ONE page accounts for 5% of the hits on edmonton.kijiji.ca? Even if there were twice as many page hits on Edmonton as the above figures indicate, this makes no sense at all.

1

Relax. Breathe.

I have sent emails like this, many times before, to online advertising providers, always in regards to new accounts for clients I wanted to impress.

Even the most responsive (in my experience AdWords) advertising networks take about a week--really 2 weeks--to find homeostasis on new campaigns...especially in new accounts.

It might spend your daily budget in an hour on the first few days, it might largely ignore language and geo restrictions.

In your case, it sounds like it might be reporting impressions on 'bot traffic.

Ignore the first week's results, things usually settle down thereafter...and that's true on every platform I have advertised on.

Also, consider the reliability of the data: impressions/views aren't very reliable (any 'bot can trigger a page load). Clicks are better, conversions are better still (especially if your conversion includes a redirect to a confirmation page).

| improve this answer | |
  • I am more concerned about the disparity between the 3 paid ads got compared to the 3 free ads. Yes, the paid aid showed my logo as well as a product picture -- both about 100 pixels. The category is currently one without much traffic, so it was easy to get all 3 paid ads and all three free ads on the same page. But a 12:1 click ratio. I don't care if it's robotic. I still had to pay for each click. Actually I do care. And I suspect them of inflating their incomes by clicking. Hence my desire to have some degree of external auditing. – Sherwood Botsford Dec 30 '15 at 15:53
  • I'm not sure of the cost but this company makes a script that tries to detects advertising fraud. One of the founders presented to a local analytics meetup I attended and he certainly knew his stuff: mktsci.com – adam-asdf Dec 31 '15 at 18:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.