Book a demo

Book a meeting with one of our marketing consultants and see how we can help you improve your ad spend and increase your LTV.

Oops! Something went wrong while submitting the form.
arrow to left
June 15, 2022

Do it in-house: Predictive Value-Based Bidding for DTCs

Maya Caspi
Head Of Product Marketing

The growth and marketing teams behind DTCs know all too well, paid media advertising is pretty much always in a state of evolution. Between rises in CAC, privacy restrictions, deprecations, and more—marketers have to put much more thought into their growth strategies to ensure that they get the most value out of every dollar spent. At the end of the day, the ultimate goal here is to maximize impact, scalability, and profitability for your brand. And this is exactly where value-based bidding (on Google campaigns) comes in. 

The thing is, subscription DTC brands have a lot more to juggle than most, considering multiple subscription packages, prices, and plans. Should any and every potential customer be deemed as equally valuable? Well, most DTC brands would work off of the simplistic assumption that this is indeed the case. In reality, if we examine the actual LTV/retention of different customers, we find they are quite widely distributed (with an equal chance, for example, of being worth exactly $X or $3X+ by, say, Month 6. As such, when you decide to apply value-based bidding in your campaigns, you bid higher for users that are worth more, and less for users that are worth less. It’s only fair to base your budget on the true value of each customer, rather than the value of all your customers.

This is not a new capability (also called targetROAS or Value Optimization) - but without predictive modeling, the value of all of the users seems the same, in the first few days of the conversion window.

So let’s dive into the inner workings of value-based bidding for DTCs, how you can apply it in-house, and further amplify its impact for the long term.


Quick explanation of value-based bidding

Value-based bidding (VBB), which is also known as “smart bidding” refers to some of the strategies used to let Google automate your bidding. tROAS, tCPA, Max Conversions and Max Conversion Value are the options there, and tROAS and Max Conversion Value are the ones that activate VBB. It is also worth noting that tROAS is also dubbed "Value-Based bidding" by Google. As I mentioned earlier, this is done by placing special focus on each customer, and placing greater emphasis (which is proportional to their value) on the specific customers that are more likely to be the most profitable. When compared to campaigns without value-based bidding, the ROAS on campaigns stemming from value-based bidding is significantly higher.

As you can expect, when using Google Ads for value-based bidding, Google gains a better understanding of your customer’s value, and in turn manages the bidding in a differentiated stratified approach, with higher bids on higher “bands” of customer values, and lower bids for lower value customers. This creates a much more efficient strategy - using the same given budget, to provide much better efficiency (ROI and ROAS).

Here’s why you should use value-based bidding

Value-based bidding has been around for a while, and has only recently become mainstream due to:

  1. The complications being thrown into the growth realm make it harder to operate at the same profitable scale, and we need a new super-power to maintain the numbers we could achieve a few years ago.
  2. The advancement in VBB algorithms in platforms such as Google - which, in the last couple of years suddenly exploded in their performance vs tCPA

When you use value-based bidding, you are prioritizing and enforcing more lucrative conversions, and deploying your marketing capital in the most efficient way.

At this point, you might be wondering what’s wrong with treating all customers the same. More is more, right? To be honest, without value-based bidding, it boils down to the fact that you are over-investing on lower-quality customers, while under-investing on those that are inclined to be more profitable. When segmentation takes place based on customer value, the returns become maximized, as seen below.

Now let’s go over the most valuable element of value-based bidding with an example:

Let’s assume your company has three typical customer types, each of which manifests different values to your company. These values can be due to LTV, plan price or probability to convert.

  • Customer 1 - is worth $100

  • Customer 2 - is worth $300

  • Customer 3 - is worth $450

Instead of paying the same amount for Customer 1 as Customer 3, even though Customer 3 is worth four times more, you can bid differently for each of those customers, in accordance to their potential value, while also prioritizing users that will manifest higher value.

The use cases for DTC subscription brands to start using value bidding include different price levels to differentiate products, different plans (monthly/yearly) and of course, if your subscription brand offers eCommerce top-ups.

Now let’s talk about how you can get started with value-based bidding, so you can make the most of your marketing budget to target higher value customers.

How to get started with value based bidding for your DTC in-house

At this point, value-based bidding is for most every brand, and here are the three steps to get started:

Step 1: Share better data - Conversion tracking

Start by sharing conversion values according to the box/product size and plan. For instance, if a user paid for the highest-priced box and chose a yearly plan, they are probably worth much more than a user that purchased the basic box and chose a monthly plan.

Now if you want to take this to the next level—your team can also leverage your company’s zero-party data, to better evaluate your users' loyalty and LTV.

Based on our experience working with D2C customers, the answers to simple questions implemented in onboarding questionnaires can be very strong indicators of future value. Especially when the questions are something like: “How many times a week/month do you plan to use our product?” or “What do you use our product for?” We recommend you analyze and identify the answers to such questions, as those strong indicators for high LTV.

Step 2: Assign value to data - Conversion values

The second step is to assign different values to conversions, based on your real-time and historical data. 

I’ll explain how it’s done. 

Let’s say you want to calculate the value of a user who purchased your basic monthly subscription.

To start you'll need to know and calculate the average loyalty of users who purchased a monthly subscription. Is it three months? Perhaps six months? Or maybe 18 months? 

Next multiply that by the monthly revenue. You can also include average additional revenue from e-commerce top-ups to the basic box.

The final result is the conversion value, as you can see in the graphic below.

Of course the accuracy and effect of the value can be enhanced by leveraging LTV models that predict users future value.

There are different ways to assign different values:

  • Static values - as seen in the example above. This is the most simple way to get started, and we recommend starting with this approach if you’re new to value bidding.

  • Dynamic values - these are changing actual values based on the actual conversion value, such as how much the customer paid, or specific product SKU’s or preferences. This is recommended when there's high volatility between customers' values/products' pricing. If you have set pricing for products, using static values will be much easier.

  • Advanced dynamic values - this refers to using predictive LTV models that evaluate and predict a user’s LTV based on any and all the engagement, attribution, transaction and zero-party data. This, of course, is the most advanced and sophisticated method. Thanks to advances in martech, you can partner with a service that can assist with LTV optimization, and other elements of predictive marketing. If you’re considering going that route, here are nine questions to ask before onboarding a predictive solution.

Step 3: Transition to ROAS - Value-based smart bidding

The final step on Google Ads is to set the campaign up as a tCPA for a short period, then switch to tROAS. The tCPA phase is only needed the first time you set such a campaign, not in subsequent campaigns.

You start this by reporting your conversion value to Google Ads on an ongoing basis, four weeks before opting in. This is done by creating a new conversion on Google Ads, and applying value to the conversion. This includes choosing whether the value is constant or dynamic. 

After four weeks, you can launch a new campaign(s) on tCPA, optimizing toward this set of conversions. It’s recommended to keep the campaign as tCPA for about four weeks, or three sales cycles. Why four weeks? Well, mainly because of the learning curve. You need to give Google’s algorithm some time to process the data.

After the waiting period is over, and Google’s algorithms have collected enough data to predict user behavior, you can switch to tROAS.

Pitfalls you need to avoid

There are some pretty common pitfalls that prove to be costly for DTCs. Here are some of the top ones that we really recommend you steer clear from:

  • Aggressive ROAS goals - It’s best to start with the recommended tROAS based on historical performance so that the system gets a good baseline. From there, you can slowly change the tROAS and periodically review if these small tweaks are getting you closer to where you’d like to be. 
  • Aggressive target ROAS changes - If you update tROAS by more than a relative 20% each day, the campaign may either stifle spend, or spend too much. Find your sweet spot slowly, but surely.
  • Over-analysis during the learning period - It’s called machine learning for a reason! The system takes time to calibrate and settle in, so give it the required 1-2 weeks to do this before you start analyzing results.
  • Looking at the wrong metrics - Google has really trained advertisers to care about click-through rate, conversion rate, and even ROAS. But don’t lose sight of how those metrics relate to your business goals. 

The benefits of Predictive Bidding

Now if you’re really ready to spice things up, you might want to consider adding Predictive Bidding into the mix. After all, the ability for DTCs to send a value based on LTV is major! If your DTC business engages with customers with different values based on variables such as subscription level, upsales/cross sales, churn and more—there’s clearly a lot of volatility at play here, and predictive bidding helps bring order and direction to your campaigns.

This branch of predictive marketing takes value-based bidding to a whole new level. Let’s start at the very beginning. What is Predictive Bidding in the first place?

Predictive Bidding, in the context of growth marketing, refers to the usage of predictive models that can analyze your brands zero- and first-party data, to accurately determine each user's future LTV in real time. By extension, your team can use this to accurately gauge each user's LTV and loyalty, based on defined behaviors, and make decisions for days/months/years from now back on that day. That’s futurespection.

For marketing teams, Predictive Bidding can be considered the holy grail in the value modeling approach. Here are some of the main reasons why:

  • Machine Learning empowers marketers to further break their audiences down to the minimal size, enabling the generation of viable predictions on a team/account level
  • Granular, value-based segmentation unlocks growth
  • Ability to make better, more informed business decisions

  • Ability to scale UA, while maintaining healthy unit economics with Predictive tROAS smart bidding

The usage of Predictive Bidding will drive the best value for your campaign budget in a sustainable manner, locking in long-term profitability. 

By now, I’m sure you realize why there is no time like the present to make the switch to value-based bidding, to maximize your conversion value. And if you choose to capitalize on LTV optimization on top of that through predictive modeling—the sky's the limit because that sets your campaigns up to succeed from the onset. It is well-known that additional conversions could happen when a new customer exhibits signals they will become a high-LTV customer.

If there is one takeaway I suggest you walk away from the post with, it would be that you really should consider value-based bidding before your next ad campaign. This can help you achieve the big goals you’ve set for your business, and get on the path to long-term sustainable success.

stay updated

join our newsletter

Request Sent. We'll be in touch!
Something went wrong. Please try again.

Continue Reading


Stop Retrospecting. Start Futurespecting.

Request Sent. We'll be in touch!
Something went wrong. Please try again.