I'm working for a company that is beginning with online advertisement and the following challenge was set:

How can you monitor the effectivess of the advertisement in actual sales?

The real challenge here is measuring how many of the people who only got to know the company by the banner made a purchase in the physical store. There's no such thing as an online store but only an institutional site displaying the latest merchandise and basic informations.

Further details

  • The company doesn't sell anything online yet. The link would redirect to an institutional page.
  • The banner ad we're working on would link to their online catalog - only pictures, no purchase button.

3 Answers 3


Tracking offline conversion is tricky, but here are some potential strategies to measure the effect of an online ad on physical store sales:

Build a discount voucher system

Create a coupon code system to track the source of sales. This could be a simple offer code that visitors quote at the physical checkout, or a voucher to print. Here's a system to minimise the chance of friends sharing the code so you measure only people who've clicked the ad:

  1. The customer clicks on the ad, which uses a link including a query string such as http://yoursite.com/yourpage?source=campaignname
  2. Some code in the head of the product page detects the presence of the source query string, sets a cookie called 'source' containing the source information, and redirects immediately to: http://yoursite.com/yourpage
  3. Some code on the product page detects the presence of the cookie, and shows an HTML voucher with the following text:

"Get 10% off [the product name]. Print this voucher and present it at [store name or address]. (For merchant use only: REDEEM23A)".

This redirection system ensures that, should the user pass the link on to a friend, their friend sees no offer code (because the query string has been stripped from the URL, and the friend won't have the 'source' cookie). There's nothing to stop someone printing multiple vouchers and handing them around, of course, but it still makes the barrier a little higher.

Improve it using unique offer codes

If you wanted even more control over the offer redemption, in step 3 above, you could generate a unique code, store it in a database, and put that offer code on the voucher. Then, at the end of each month, collect up all the vouchers and check the offer codes against the ones in your database to rule out duplicate codes and build a more accurate picture of who used your vouchers.

In addition, the section that appears for customers who click the ad could prompt them for their email address and send them a unique voucher code. They might be more likely to print the voucher if its stored safely in an email.

I appreciate that your client might not be too keen to offer 10% off their products, but you could reassure them that it will likely save them money; offering a small discount will help them track the success of their campaigns to ensure that they're not wasting money on ineffective marketing.

Target the ad at one geographic location only

If your customer has multiple stores, you could run the ad only for visitors from a certain area and see if sales in that area pick up. Facebook and Google AdWords both offer geo targeting of ads.

Run only one advert

If you can't geo target your ads for any reason, you could target by time instead. Run only that ad in a given week in which sales volume is generally stable and well known, and measure effect on sales.

Turn it on and off

Run the ad for four weeks, and turn it off for four weeks and measure the difference against the usual baseline of sales for each four-week period.

Don't run online ad campaigns back-to-back

Have a small gap between online ad campaigns to see if sales return to the baseline in between campaigns, and to better gauge the difference between two ads you're trialling.

If measuring phone sales, use a different number

Detect traffic from the ad using the method described above, then show a different phone number on your page if you'd like to gauge the impact on phone sales.

In summary

It's very hard to do accurate cost-benefit analysis for online ads when you're measuring conversion rate offline, but using a custom voucher code system and carefully limiting the location, frequency, and number of ads you run at one time are among the best methods to measure ad effectiveness.

  • I updated the description. It could be giving the wrong idea the site has an ecommerce feature. I like your idea about a landing page but we're still missing a link to check who of those people who clicked the online ad finished a purchase in the store.
    – Renan
    Jul 6, 2011 at 17:30
  • Thanks; I had misunderstood your question. I've modified my answer to deal with measuring offline conversion of online ads.
    – Nick
    Jul 7, 2011 at 7:33

You can use Google Analytic's Campaign Tracking for that banner ad which will allow you to get specific information about users who click on that ad. It can't tell you who is actually buying anything since the sales occur offline but if they do eventually sell goods online that can be tracked as well.

  • That's not what I'm looking for but I appreciate your concern.
    – Renan
    Jun 9, 2011 at 16:04

Similar to TV commercials that say "Ask for agent 347" or "use promo code 219" you can ad something to your Advert that helps indicate which placement/location was the source for your lead.

If you combine this with Google Analytics (or similar) you can see where users are coming in from.

The only trick with this is that at large scale the accuracy drops. E.g. If I watch a Diggnation video and it tells me that I can get a discount on hosting at "you know who" by entering the promo code "digg4" it helps track where I came from... but now that these promo codes are "out there", many times I just go to the site to make my purchase and then Google to find out which promo code is the best for what I need (even if I've never visited the site that offers the promo code "FunWithSheep".

Finally as much as it is a desire to track this kind of info, it often fails. If I go to McDonald's was it due to the TV commercial I saw? (which one? this weeks? last weeks?) a radio ad? or just the fact I've known the menu at McDonald's for 30 years and had a craving for a Big Mac.

I'd likely trust a statement from the end customer most. Just flat out ask them where they heard about "Company/Product X" and give them a "free-text" entry textarea to tell you. They may give you great insight... e.g. not just "a friend" but rather "Jim Bob over at InnoTech told me that {YourProduct} saved him 20% of his payroll last year so I decided to check you out". This is way more helpful, since you now know that word of mouth is powerful, and that maybe touching base with Jim Bob to see if there is anything else he needs/wants... or if you can use him for a testimonial success story in future ads.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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