THE RKGBLOG

Announcing “No CSE Sales, No CSE Rows Week” — April 1 thru April April 7, 2007

Does your firm advertise on shopping comparison feeds?

If so, you’re cordially invited to participate in “No CSE Sales, No CSE Rows Week”, scheduled for April 1 to April 7, 2007.

Here’s how “No CSE Sales, No CSE Rows” works. Right at the end of February, 2007, for each CSE determine all the SKUs in that feed which had exactly zero corresponding sales in February. For one week, April 1 through 7, drop these items from your feed for the corresponding CSE. In short, for each CSE, remove their zero-sales items for one week.

Why do this? What will happen?

As these are items that aren’t selling, your sales shouldn’t decrease. And as these items that aren’t selling, this change shouldn’t provide any advantage to your competitors. Projected sales and share impact: nil.

What about costs? If these items aren’t generating sales, you’d hope they’re also not generating click fees. If that’s the happy situation, then dropping these SKUs for a week shouldn’t impact costs either. Projected cost and earnings impact: nil.

Why participate in a project with no anticipated top or bottom line impact?

Two reasons.

First, what if costs ,instead of staying flat, actually decreased during that week? That’d suggest the no-sales items were consuming cost and were reducing your profits. Hmmm. That’d indeed be an interesting and profitable thing to discover.

Second, do you know today what would happen if you did this experiment in April? Does this proposed experiment leave you feeling a bit uneasy or uncertain? If you don’t know what would happen, or it the experiment does make you feel uneasy, perhaps that’s because you lack sufficient insight into the profitability of your CSE advertising at the SKU level? If so, don’t blame your marketing team. Blame the opacity of the feed engines when it comes to providing “atomic” (eg, by SKU by day) cost data.

Here’s another thought experiment. Take your top 10 performing Google keywords and contemplate turning them off for one week. Crazy, you say? Indeed. You could likely tell me in short order the cost of those words and their corresponding sales. And I bet they’re profitable. And so I bet you’d be crazy to even think of turning them off.

I’m not advising you tinker with the successful portions of your Google campaigns. I offer the Google example to contrast with the feed situation. Google (and Yahoo, and MSN) provide advertisers with detailed atomic cost data both through their web tools and through their APIs. Kudos to them for doing so. Cost data lets direct marketers manage their advertising intelligently. The search engines know that and support that.

Not so for most comparison shopping engines. For the most part, cost data isn’t easily available to advertisers — that is, at the by-day by-SKU level. Certainly not via APIs or through machine-friendly formats. (And no, screen-scraping costs isn’t acceptable in 2007. Yes, it works, but is a tremendous IT pain for all sides.)

CSEs should provide their advertisers with detailed cost data and they should provide it in machine-friendly form. Lack of CSE cost data impedes direct marketers from managing their advertising intelligently.

So, again, you are cordially invited to participate in “No CSE Sales, No CSE Rows” week. Think about it. And if you’re not sure you want to participate because you’re unsure of the impact on your business, well, then I’ve made my point.

Join us and the ARTS in the effort to standardize data formats for getting data into and out of comparison shopping engines. Welcome!

Technorati Tags: , , , , ,

  • Alan Rimm-Kaufman
    Alan Rimm-Kaufman founded the Rimm-Kaufman Group...
  • Comments
    5 Responses to “Announcing “No CSE Sales, No CSE Rows Week” — April 1 thru April April 7, 2007”
    1. trent says:

      Your post is somewhat myopic and assumptive suggesting that the reason intelligent marketing can’t happen at the sku level is because they don’t have the data. It is true there are CSEs that don’t provide the sku level data yet some do and the problem of calculating sku level ROI still exists. Why? The more likely culprit is that lack of functionality within the ever important technical system to manage all the “atomic” data. In fact, most retailers that I’ve spoken with have told me that “I would love to have all the data for every product BUT I wouldn’t be able to do anything with it because we don’t have a way to process it.” I’m not suggesting that the data isn’t invaluable because it is but you need a car before you can put gas into it.

      Thus the suggestion to blindly remove products from datafeeds doesn’t make any sense and is irresponsible if retailers can’t track product sales by channel in the first place. Additionally, just because products don’t get sales doesn’t mean that they aren’t contributing. Some skus might be considered a “loss leader” of sorts and while not having any direct sales for that particular product but may be contributing to sales of another. That attribution is most likely not being accounted for either and for that reason shouldn’t be removed from the datafeed.

      After April Fools Day, I propose that the first half of April be dedicated to investing in data feed management systems so that retailers can really add or remove the right product, at the right place, at the right time.

      /trent

    2. Hi Trent –

      You wrote: “…the suggestion to blindly remove products from datafeeds doesn’t make any sense and is irresponsible if retailers can’t track product sales by channel in the first place….”

      100% agree. If retailers can’t track sales by channel, they have far bigger problems.

      Any decent web analytics package, software system, or agency can provide detailed sales tracking. To extend your car analogy (eg “you need a car before you put gas in it”), I’d say if you can’t walk across the driveway to reach your car, you’re in pretty rough shape and likely shouldn’t be driving and may not be ready for either a car or for gas. Tracking sales by channel seems (imho) as fundamental to direct marketing as walking is getting around. And about as basic.

      (Now all tracking systems have limitations — certainly. Channel spillover, competing cookies, etc etc. But my sense wasn’t that you were suggesting those advanced challenges were what were impeding retailers — my sense was that you were suggesting that basic sales-by-channel approximate tracking was difficult. Apologies if I misread your intent.)

      You wrote: “…the suggestion to blindly remove products from datafeeds doesn’t make any sense and is irresponsible…”

      100% agree. Retailers shouldn’t make changes if they don’t understand their impact.

      I picked April 1st as this suggestion was partially tounge-in-cheek.

      But, pulling the thread, why are retailers (to use your word) “blind”? For those retailers who are “blind” because they lack solid tracking systems, fine, they can choose to invest in some software, they can pay an agency, whatever. No worries. What gets me is the portion of “blindness” due to obfuscation (intentional or not, that’s not for me to say) on the part of many of the feed engines — they don’t make readily available the data that serious direct marketers need to manage their feed campaigns

      You wrote: “…while not having any direct sales for that particular product but may be contributing to sales of another…”

      Any reasonable tracking system should be able to track both direct and ancillary sales. I’d agree with you 100% had you written something like “… while not having any direct sales for that particular product but may generating significant sales in stores or into the call center”. I don’t agree that a feed ad for SKU ‘x’ that generates a sale on SKU ‘y’ poses any challenge to a robust tracking system.

      You wrote: “…I propose that the first half of April be dedicated to investing in data feed management systems so that retailers can really add or remove the right product, at the right place, at the right time…”

      When you wrote “at the right place, at the right time”, do you mean ‘by instinct’, or do you mean ‘based on the economics?’ If the later, we’re back at the cost issue.

      A fun and important debate, thanks!

      Cheers –

      Alan

    3. trent says:

      We’re in violent agreement. :)

      cheers.

      /trent

    4. The idea of optimizing your data feed by removing under performing products (SKUs) has always been a topic of debate. On the one hand you have the camp that adamantly says removing products is always bad. On the other side you have those that believe removing your under performing products IS the way to optimize your feed.

      I feel that removing under performing products is part of optimizing your feed. Why pay for un-useful clicks? And any marginal benefit (sale of a different item, branding, foot traffic) you might get from those clicks on under performing products is probably still not beneficial enough to warrant paying for clicks on it over the course of the year. Removing under performing engines is also part of optimizing your feeds, and to a greater extent, removing products only from certain engines they do not convert on is what should be done. But yes, this is ALOT of work. That is why most merchants do not do or they do it ineffectively. Using and experienced agency or consultant is probably the best thing for most merchants.

      Hey, I love the site Alan! Keep up the good work.

    5. search_junkie says:

      I forgot to mention that one of the best ideas I have heard of is to remove under performing products throughout the year and then if things get stagnant, throw them back in the feed. And especially during the holiday period you will want to submit your entire product line!