Cascade retargeting strategy (4 deadly sins checklist included)
Cascade retargeting strategy is discussed, along with what to do if the retargeting lists are too small. A not so brief overview if common pitfalls included
'Cascade retargeting strategy (4 deadly sins checklist included)'
Cascade retargeting strategy (4 deadly sins checklist included)
Retargeting is just a tool, so I’d rather discuss a retargeting strategy called “Cascade”.
Note, that in context of this article retargeting equals remarketing.
What’s the difference between retargeting and remarketing?
Words ‘retargeting’ and ‘remarketing’ often used interchangeably. There is a small difference only in terms of delivery method:
- Retargeting is based on cookies (e.g. Facebook ads and other social media as Twitter, LinkedIn, Instagram)
- while remarketing is based on email list (e.g. Google Display Ads)
Retargeting and remarketing are both effective methods in their own right, yet a combination of both may be the best strategy to boost your digital marketing strategy results.
Does retargeting work?
Some case studies claim the conversion rate for potential customers who have been retargeted is 70% higher than those who haven’t. The click-through rate (CTR) of a retargeted ad is 10 times higher than the CTR of a typical display ad and so on.
Retargeting use cases
Here are some patterns of its use off the top of my head:
- Pre-sale: when a website visitor has not committed the target action yet. Often a decision to purchase is not immediate. It makes sense to remind customers of your offer, while they are in the heat of the decision making.
- Cross-selling, upselling, etc.: when a user has already been “converted”. You can offer some additional services (a dedicated landing page is a must here), before the user forgets about you and walks away. Or another example: say a user was looking for an excerpt from the set of real estate market laws: maybe, in the near future, the user will need the services of a real estate agent or a lawyer to buy property.
- Refinement of advertising. Example: a user initially came to the website through a request of “legal services in Boulder”. And while on the website a user was looking through the information on the inheritance laws. You can commence classic remarketing or “search remarketing” — Google AdWords RLSA — to specifically offer inheritance related legal services.
These are just the most obvious ways.
Usually, retargeting is simply applied to all website visitors, which did not perform a desired action (call, request, etc.). But at different stages of decision making (i.e. marketing funnel), a user needs different information.
Retargeting strategy: from universal to cascaded
Let’s pretend that the timeframe of decision making in our niche is 7 days average.
If the user has been on the website recently (say, up to 3 days ago), through requests with words like “urgent/fast”, the user is totally ready to perform the target action (purchase). It makes sense to offer such users some kind of an additional bonus (such as free shipping) to come back and complete the purchase.
In similar manner, we can assume that the user who was on the website up to 7 days ago, is less likely to convert. The user probably analyses the market, compares offers and reads reviews. So we should tell this user about additional advantages of our offer, which positively distinguish us from competitors.
As 7 days went by, the user has already executed the target action (if not on our website, then somewhere else). In this a situation, the user no longer needs what we originally proposed. But hope is not lost. If the promoted product/service has a pack of additional, related goods/services or accessories to go along with it, it’s time to offer them to our retargeting database, even if they already purchased the product itself, even if they bought it from our competition. So we work 7 more days after the sale (up to 14 days from the first visit to the site).
If consumers buy an advertised product/service periodically (once a month, once a quarter, regularly), it is advisable to try to “activate” the audience again after some time. We already know that our niche is interesting to these users, because they have already been on our product page.
A good example will be construction companies. They purchase construction materials on a regular/periodic basis.
Let’s make a retargeting list (custom audience) with a participation period of 14 to 60 days. Now, instead of one common list, we got four. They are sorted by participation period:
- Users who were on the website up to 3 days ago
- Users who were on the website from 3 to 7 days ago
- Users who were on the website from 7 to 14 days ago
- Users who were on the website from 14 to 60 days ago
All these users were on our website in the past and showed their interest, but each of the 4 groups requires a different approach and a different offer. Groups of users with different participation periods form a cascade, and the retargeting becomes cascaded.
To make a retargeting group with a participation period of 7 to 14 days, you must first compile 2 lists of audiences with a participation period of up to 7 and up to 14 days and then create a combination. Subtract the first list (up to 7 days) from the second list (up to 14 days). The resulting list will consist of users who were on our website from 7 to 14 days ago.
The Method of 7 reasons
It’s a variation of cascaded retargeting strategy. We, just like before, create several retargeting lists with different terms of participation. For example: 7 lists, 1 day long each.
We make up an individual Unique Selling Proposition (USP) for ads for each list with one common idea: “This is your first reason to choose our company”, “This is the second reason”, etc. Works better for specific products and services rather than “pick our company because…”
What if the remarketing lists are small?
AdWords packs some inherent limitations on the size of the list:
- for Google Display Network (display ads) it is a minimum of 100 users on the list over the last 30 days
- for the search remarketing RLSA it is at least 1,000 users over 30 days
Here is a link to AdWords requirements.
Even if the minimum threshold is reached, there may be not much users there to justify retargeting efforts. A few users in the list — means only a few returning website visitors. The time spent on remarketing campaigns may not be justified, in that case.
That’s when Google recommends expanding the “remarketing window”. For example, collect the list of people with a 60-day participation term, instead of 30. On one hand, the list seems to be expanding, but the main point of the remarketing tool itself is lost — the precious time to persuade people to make a purchase decision. Say you will catch up with people who were on the website 60 days ago. Then what? They have already forgotten what they were looking for, what they were interested in and they don’t remember your company at all (so that equals to noisy cold outreach again). And what if the time-window to make a decision in your business is not 30 days (as by default), but 14 or even 7 days?
Don’t give up your retargeting efforts yet, just keep on reading.
“Similar audiences” in Google AdWords remarketing
Use another Google tool called “Similar Audiences” along with the principle of positive feedback from circuit design.
Google continuously analyzes users behavior on the Internet, taking into account a huge number of parameters (visited sites, search history, usage frequency, preferences, etc.). Thanks to machine learning, “similar” people are easily defined among all users. Google looks for common signs that allow you to form interest groups.
Let us assume that the timeframe for making a purchase decision in our business is 14 days (the most effective timeframe to use remarketing).
Step 1. Compile 2 remarketing lists
- With a participation period of 14 days (our work list).
- With a much longer period to accumulate as many users as possible. Let’s make it 90 days.
Fine. Wait for the lists to fill up.
After some time (depends on how fast the “longer list” will be filled), Google will automatically generate a list of users “similar” to users in the “longer” list.
The main advantage is that a user does not even have to visit any specific page on our website in order to get into the “similar” list.
What makes it work is that the behaviour of such a user on the internet is “similar” to the behaviour of people from the “longer” list. Unsurprisingly, people with similar behaviour often have similar interests.
Step 2. Launching the advertising campaign in the Google Display Network for a “similar” audience
We create a classic advertising campaign for the GDN, but target a “similar” audience. We get a significant increase in targeted traffic to the advertised service and at this point work on brand awareness rather than selling anything specific. That’s when a content marketing can be very handy, because referring users to educational \ provocative \ interesting content results in lower bounce rate.
At the same time, the “shorter” remarketing list is also expanding.
Step 3. A “combat” advertising campaign for the “shorter” remarketing list
The audience on the “shorter” list has expanded. Now you can use remarketing to its fullest (e.g. using Cascade strategy we talked earlier).
The positive feedback is: the more people are accumulated in the “longer” remarketing list, the bigger “similar” audiences will be. In turn, it increases traffic to the advertised service or product and the size of the “shorter” list.
In addition it simultaneously expands the “longer” list. So it is an infinite loop 🙂
Deadly sins of retargeting
1. Incorrect Segmentation
Incorrect segmentation belongs to one of the most common marketing mistakes. The analysis of users’ behavior on the advertiser’s website and their shopping intentions is a key to determining a retargeting campaign success.
Advertising channels, where users come from or where they’re going to see a retargeting ad are of minor priority. Our interest heavily drifts towards each individual user’s conversion ability. We need a buyer at the end of the day.
However, marketers often try to segment retargeting campaigns by the source of first time visit. They suppose that those who come from context ads are converted better than those from organic search. While this is generally might be the case if you look at Google Analytics numbers – in reality it is more about about conversion attribution model rather than retargeting strategy.
Following the same logic advertisers want to control platforms of retargeting: choose platforms in the first place and manually shut down ineffective ones, or even shut down the whole channels. However, what you really need to do is shut down ineffective users, not platforms. Marketing efforts targeted at wrong auditory go in vain, but figuring out a right auditory is a try and error game itself.
Another thing is trying to divide retargeting contractors according to advertising tools they have in order to minimise the competition between them. As a result the efficiency of each of the contractors goes down and users get overwhelmed by ad messages. That happens because every contractor is going to compensate the lower coverage by higher view rate.
All those segmentation examples from above contradict the very concept of retargeting: showing ads to specific users on some 3rd party platforms (e.g. target those who visited your pricing page on social media). We need to show them a banner carrying the right message, which was formed based on the analysis of user’s interests. We need to show the banner at the right time, individually limiting the number of times it’s displayed, determining an individual specific price for each view.
Do not rush into extremes though. Telling users who interacted with your brand already (or even better – current customers) about your new product line is also a viable retargeting strategy.
Shopping behaviour and purchase history
You might also want to consider the shopping behaviour and purchase history of users. The first step is dividing them into groups of old buyers and new buyers. The majority of online stores value “new buyers”, who haven’t purchased anything yet and companies are ready to pay more for these new users.
But the problem goes a bit deeper than it may seem. Say you get a new buyer, but how do you know what one is up to do next? What the average order value will be? How many times this buyer is going to shop at your store?And now if you have a loyal customer, how can you tell this customer is not going to just leave you one day, switching to your competition?
Our job is to not just gain a new customer, but convert one into a loyal customer and sustain the loyalty over the whole customer lifecycle via multiple touchpoints.
We can do this by segmenting the audience using CRM data: evaluate the value of each segment and adjust advertising investments, focusing on segment of the most interest.
We can, for example, segment users that had no purchases for a long time already and invest our money and effort into restoring their loyalty. You can also spend more money to bring back specifically those, who generate the most profit for you.
And it’s not just about having different Cost Per Acquisition (CPA) values for each segment. It’s about a possibly radical change in communication strategy: you can have different ad banners for each segment, holding different offers with different messages.
Let me give you an example of how BrandAlley (shopping-club) managed the segmentation process. They were able to increased ad campaign ROI to a previously seemingly unreachable level.
- First, the audience was segmented into new customers and existing customers. Based on eCPA (effective CPA) for each segment and BrandAlley put most of its efforts into new customers to increase conversion.
- The second step was using the data on user purchase history from shop’s CRM system (btw, for flexibility in such efforts GRIN tech advocates for custom CRM systems. Out of box solutions works well for idea validation and early growth stages)
- Now, as soon as a user visits any page on the website, his profile is compared with CRM data.
As a result it was possible to identify 80% of the audience and segment it, according to their purchase history. They later set up an ad campaign accordingly, investing more funds into segments that routinely give more profit or are at risk of losing their loyalty.
This exact approach allowed BrandAlley to focus on the most valuable user segments, regardless whether they’re newcomers or old customers. This ad campaign brought BrandAlley four times more sales and two times more customers.
2. Overwhelming intrusiveness
Overwhelming intrusiveness is another common fallacy in retargeting strategy implementation. Retargeting can be very annoying and it’s the main reason users dislike it. Advertisers who value their reputation pay attention to such things.
Showing banners to people who have already made a purchase, showing multiple ads on one page (or just spamming a Facebook feed) or having no limit to how many times banners are displayed (banner exposure frequency cap) is not just annoying to users it also is an ineffective use of a budget.
How can you keep your brand’s reputation and increase budget effectiveness? Here is some quick thoughts:
- Once a user makes a purchase, stop showing banners — at least of the same product.
- Use different offers for different audience segments (showing a relevant ads sounds like a no brainer, huh?)
- Implement deep ad personalisation e.g. taking purchase history into account.
- Stimulate demand.
- And one more thing: only show one banner on a page (Ad Exchange feature). Be smart about limiting the frequency of exposure — setting up an exposure cap to 30000 views is not enough. Spread these over segments of time: set a limit per hour/day/session. It’ll help you avoid bombarding users with ads when they’re least responsive to them.
Placing an ad directly on a certain platform or Facebook allows to limit exposure only for an individual channel.
Thus, a user who already saw the ad a required number of times (say on Facebook) faces it again once redirected to a website, which is included in Google Ad network. Now multiply it all by the number of digital devices the user has: work laptop, home desktop, smartphone, tablet. So we basically have no real control over banner exposure frequency cap to a single specific user. Real-Time Bidding technology is what’s might help you with smart exposure limiting.
The time of banner exposure is one of the key factors in retargeting. Smart limiting of exposure intensity allows you to control who, when and how many times sees your ad banner.
3. Wrong sales attribution model
Sales attribution is a method of assigning conversion to marketing channels.
Wrong sales attribution model or a series of typical mistakes can lead to a decline of buying activity across all paid marketing channels.
The vast majority of marketers cluelessly use Last Click or Last Cookie Wins model. Conversion value in such model is assigned to the last channel a user came from.
Consider the following scenario: a user visit a website after seeing a context ad and clicking it. Later this same user visits the website again, but now it’s through an affiliate program. And for the third time the same user visits the website again after clicking a retargeting banner and finally makes a purchase. Who gets the sales point? Retargeting campaign team (duh).
In reality the sequence is way more complex and this example was ofcourse way too simplified. However, only a fraction of users make a purchase (or any specific action you want them to) after the first ad click. Users return to a website multiple times and every time they usually come from a different source. You cannot affect these factors directly, so it is important to know what to consider and why.
- First of all, you need to exclude direct user requests like when they type in your web address in a browser. These are a direct consequence of your advertising activity. People can’t type your web address just out of blue.
- You need to exclude brand-requests for the same reason. People type in an address of a store (or its name) not in the address bar of a browser, but in the search request field on Google. They can visit your website through an ad banner, as well as clicking on link in the list of search results.
- Hypothetically, advertisers should exclude all the search traffic, since it is not a paid channel.
- Exclude email listings on top of that. You can’t increase the mailing budget and boost the number of orders this channel gives you. The number of your mailing list subscribers is, once again, a direct consequence of your advertising activity.
The most important feature of any paid channel is the ability to increase volume with increasing budget. Once you see that a particular channel is effective, don’t hesitate to invest more and get ready to reap more orders.
Thus, investing in “free” channels decreases overall buying capacity of your paid channels, including retargeting. (Editor’s note: should we then shut down our content marketing department? 🙂
As we can see the “last click” model is not perfect: users interact with many channels before making a decision and only one channel is rewarded.
Advanced marketers use “Customer Journey” model to pay heed to a contribution of each channel to sales.
Within this model a sales reward is allocated between all links of the chain. But there contribution wighted differently. For example: 40% goes to a channel, which brought a user to your website and 40% goes to a channel which brought the user back with a following purchase. The rest goes to other channels the user was interacting in between those two events.
That is how you can measure a level of contribution of each channel to sales and allocate your advertising budget more efficiently.
4. Ignoring Real Time Bidding
For the last couple of years Real-Time Bidding has been one of the hottest topics in advertising. RTB and retargeting are often so closely related that many people confuse these to concepts.
I’d like to point out once again – RTB is a cool tool for businesses of certain size. For an idea validation and early stage companies this is not something worth chasing.
RTB allows advertisers to not just choose, but to individually evaluate the most profitable views. You can significantly increase the effectiveness of dynamic retargeting by comparing the cost of a single view per user to the probability of the user’s conversion.
The price that individually depends on which user, for how many times and where sees an ad is a key part of a retargeting success.
We at Sociomantic consider up to 30 various criterias to calculate the appropriate price and use RTB to its full potential. We make up to 16000 different biddings, while most make from 4 to 15 typical ones.
But RTB advantages are not always accessible to an advertiser. When contractors track PPC and target actions, they can use RTB to increase their margin.
Let’s say an action cost is established at $10. With the help of RTB a contractor get it for $2 and gets a margin of $8, which goes straight in the contractor’s pocket. Such a model implies that a contractor has a goal of buying as cheap as possible but still trading at a previously agreed price.
The alternative is to have a contractors premium fixed and stated in a contract from the start. Now a contractor gets to keep only a certain percentage of money and the rest will be reinvested again. Now the only way for a contractor to earn more is to maximize the number of sales according to CPA.
This transparent model is the only way for an advertiser to gain full advantage over RTB and interact with a contractor upon KPIs only.
To sum it up on retargeting strategy
When I was a kid there was one strong association with advertising industry — creativity. That’s what was so attractive to me about advertising career. To this day the industry attracts huge numbers of creative people full of ideas and inspiration.
However, the world has changed.
Online advertising leaves less and less room for creativity and makes manual operation almost impossible. In the modern world effective ways of implementing marketing strategy shift as humans have no chance against the power of digital technologies (but can utilise them to a great benefit).
It is not a marketer’s creativity that defines the efficiency of an ad campaign, but an instant analysis of huge piles of data, lightning fast decision making and real time adaptation.
This is the only way to use retargeting to its full capacity.
If you’d like GRIN tech’s marketing team help you execute a retargeting strategy or just have a chat whether it is a fit at all – do not hesitate to get in touch.