While this is a valid concern, there are a few ways you can limit your creep-quotient, including diligently researching and applying negative audiences to your campaigns, using burn pixels and employing frequency caps.
The first 2 are relatively easy, they only require thoughtful "thinking ahead" when building out your campaign. Frequency caps, on the other hand, are not so straightforward. How much is too much when it comes to remarketing?
Is there a magic number of exposures?
I've seen figures like "maximum 3 exposures per unique visitor per day," "7-12 exposures total," and even "1 exposure per day." Support for these recommendations are often "advertising industry best practice," which may be true for general display -- but is remarketing really the same as display advertising?
Whether you are site or search retargeting, your remarketing ads are targeted to web users who have expressed intent, either by previously visiting your site or querying a relevant keyword in a search engine. This enables you to build campaigns that are more relevant to your targets, including creative that is personalized to contain items left in a cart, search terms performed on a site or cross-sells based on items already purchased.
Turning up frequency can turn up sales
An example is Indochino, who began with a 3-exposures-per-unique-per-day approach. Bumping the frequency cap up by 3 in every test, Indochino kept seeing revenue increasing. "We got as high as 15-20+ impressions per day per user. Every time we've increases it, we've seen greater revenue and ROI. Users don't seem to mind," says Omar Al-Hajjar, Online Marketing Manager.
If Indochino did not question the conventional wisdom of "3 exposures per day," its campaign would be driving only a fraction of the sales it currently does.
But sales are not the only benefit a high frequency cap has rewarded the company. High exposure makes the company appear "bigger" to web users who are not aware of retargeting - all of a sudden, this brand is everywhere. Its also landed business development deals because partners have seen the brand appear on sites like the New York Times, a display buy which is usually cost prohibitive to small business through such publications' first-party advertising options.
Consider your industry
Indochino sells custom made suits. To order, a customer must submit detailed measurements. Its average days to purchase is 28 days, longer than most other apparel sites. "Retargeting keeps us top of mind during those 28 days as a customer goes out and researches reviews on Indochino, talks to a buddy about which style of suit he wants, and ultimately finds a tape measure and a spare weekend to get measured at home."
How do you determine optimal frequency for retargeting?
Your testing capabilities may depend on your vendor, but ideally a true A/B split test that involves one group receiving a low exposure, such as 3 per day, versus higher frequencies would be run during a single test period, rather than sequentially.
An alternative is to create 2 campaigns with similar audience attributes, such as visitors to product category A and category B. Choose categories or landing pages that receive similar traffic, and have similar days to purchase or visits to purchase, as reported in your web analytics. Ensure these test audiences become negative audiences for your general campaigns, and that visitors who view both categories A and B are excluded from your test (they can still be members of your general audience).
It's a good idea to test the types of campaigns you run - general campaigns may bear more or less frequency than your more targeted ones, as do certain categories.
You may not find the same results as Indochino, a tighter cap may be more effective for you. But don't rely on "industry says this," or you will sell your campaign short.