By Konstantin Bayandin, Founder & CEO at

Each advertiser is aware of the worth of utilizing microsegmentation to focus on audiences almost certainly to purchase a product. However solely the savviest on-line advertisers know the inherent downfall of that strategy – and learn how to overcome it.

With predictive optimization, you’ll be able to nonetheless goal the highest-performing viewers segments with out ignoring the vast majority of your potential clients for higher return on advert spend (ROAS).

The issue with microsegmentation

Microsegmentation is a binary, all-or-nothing technique. Advertisers can use instruments akin to Facebook lookalike audiences or Google Ads Customer Match to focus on customers that resemble earlier clients. Or advertisers can target market members based mostly on private attributes, pursuits and demographics utilizing instruments like Google Ads Audience Targeting. Both an viewers meets your segmentation guidelines or it doesn’t. Sadly, there are lots of people who could be fascinated with shopping for your product – if solely they’d heard about it.

What number of clients are you lacking out on with microsegmentation?

Microsegmentation, or microtargeting, causes you to lose out on 50 to 90 % of your whole potential buyer base. Let’s stroll by means of an instance utilizing actual consumer knowledge to point out you the way the numbers stack up.

Let’s bin the shopper’s total potential viewers into 100 buckets, every representing one % of the overall viewers inhabitants. The buckets might be binned by attributes akin to predicted lifetime worth of a contact (Determine 1).

Determine 1:

Every 1% bin is sorted randomly alongside the x-axis. Conversion charge is on the y-axis.

On this instance, the most effective bin converts at 48 %, the second-best bin converts round 36 %, and so forth. Trying on the cumulative order knowledge (Determine 2), we see that the highest one % of the viewers is 8.7 instances extra more likely to buy than the common viewers member. Moreover, the highest 10 % of the viewers offers 43 % of the orders.

Determine 2:


With microsegmentation, advertisers would possibly create a lookalike viewers on Fb utilizing the highest two % of the viewers, and even the highest 10 %, however they’re actively turning away as much as 57 % of the potential orders!

Why is predictive optimization more practical?

With predictive optimization, advertisers can use conversion knowledge for every viewers phase to achieve a complete potential buyer base with out compromising ROAS just by bidding much less for lower-converting viewers segments.

Predictive optimization might be applied by way of an easy-to-follow, repeatable six-step course of:

  1. Web site pixel – A pixel is put in on a consumer’s web site to gather first-party customer knowledge, together with on-site behavioral monitoring, for each customer over a delegated time interval.
  2. Buyer knowledge – Web site habits is mixed with the consumer’s buyer relationship administration (CRM) gross sales knowledge over the identical time interval to get an entire image of buyer habits.
  3. Machine studying – The shopper knowledge is used to coach machine studying fashions.
  4. Optimization sign – The mannequin then outputs the likelihood of a person customer’s probability to transform.
  5. Good bidding – Lastly, the optimization sign knowledge for 5 to 10 % of the best engaged web site guests is fed into Google’s and Fb’s good bidding algorithm to bid the suitable greenback quantities for the suitable clients — with none focusing on.
  6. Revenue – The consumer sees excessive ROI with out omitting a big phase of potential clients.

Predictive optimization is based on the premise that each viewers phase is worth it, so long as it may be marketed to on the proper worth per impression or click on.

Why is now the best time for predictive optimization?

1. Advert platforms are smarter.

Advert platforms like Fb and Google have shifted to extra automated bidding and smart-bidding methods.

As a result of advertisers can use predictive optimization that integrates buyer habits with CRM gross sales knowledge to coach smart-bidding algorithms, there’s loads of knowledge to optimize for value-based indicators like conversions or income as an alternative of blanket segmentation.

2. Advert platforms are extra open to conversion API integrations.

As lately as final 12 months, Fb allowed server-to-server knowledge integrations just for their high hundred or so advertisers. Now, it’s open to any advertiser, enabling predictive optimization suppliers akin to to combine consumer knowledge immediately into advert campaigns for smarter bidding methods and optimized ROAS.

Predictive smart-bidding optimization makes use of predicted buy worth because the optimization purpose, so you’ll be able to ship promoting algorithms with a non-binary enter that enables advert platforms to optimize for the suitable sign – for superior conversion worth for the best engaged potential clients.

And that results in extra gross sales and optimum ROAS.