This week, we launched the Programmai web app.
The brand new interface is designed to guide you through some marketing game-changers: based on first party data and real performance.
The three key product features:
Connect is about data going in and out. We connect to your Analytics, CRM and data warehouses to bring first party data in, to join and prepare it for cutting-edge machine learning models and deep learning techniques.
Predict provides a snapshot into your future customer. We predict who’s most likely to buy your product today and how valuable they’ll be to your business tomorrow.
Activate is where you take action. From launching experiments to creating predictive audiences. If you want to improve the performance of your acquisition campaigns, this is a great place to start.
So, let’s begin there.
Here’s how you can use the Programmai app to build a powerful predictive audience in a few clicks – and in doing so, double your volume while halving the CPA.
Time to look forward.
As marketers, we use historical data to make decisions about the future. Why? Because for a long time, that’s all that’s been available. While ‘History doesn’t repeat itself, but it often rhymes’ (Mark Twain), it’s time to analyse the present in order to predict the future.
If you’re building a new audience for customer acquisition, most ad platforms use a pixel-based lookalike strategy. You’ll likely be targeting prospects based on one metric, such as a page visit or conversion.
If your objective is to drive sales, that broadly makes sense – right?
Well, not quite.
This prospecting strategy brings in customers. But there’s no way of knowing if they’ll buy your lowest value item or worse, cancel and return their order later. The pixel-based strategy drives up business costs and risks damaging your brand in the long term – effectively, trading in your customer’s future value for the next short-term purchase.
So, what are you suggesting?
Rather than relying on historic data and ad platforms for your prospecting, build lookalike audiences based on the complex interactions discovered within your first party data. By moving away from transactional data-points, you can develop a clearer holistic picture of your prospect.
Our machine learning models are designed to predict your prospect’s predicted Customer Lifetime Value (pCLV) so that when you configure your seed, you can target more of the people who behave like your best customers (before they became your best!)
In Programmai, we call these Seed Audiences (we’ll come to this more in a second).
We refresh seed audiences every three to five days. On average, we see a 40% change week on week, meaning we move with you and the seasons; always rebuilding your lookalike to attract the right customers into your brand. Never fatiguing.
Plus, everything you see here works without third-party cookies or identifiers for advertisers (IDFAs). That means when the cookie depreciates and other IFDAs are phased out, you’ll not only be unaffected by the changes and escape the competition. Crucially, you’ll be activating against predictive audiences that are proven to drive volume, efficiency and uplift. No trades-off.
How to build predictive audiences in seven steps
1. Select your objective.
You can build predictive audiences for any marketing objective. Let’s say you’re focusing on customer acquisition – with an eye also on retention. Choose ‘Acquisition’ to get started.
2. Select targeting set.
3. Choose a platform.
We use your first party data to predict individual Customer Lifetime Values (pCLV). We supply that information dynamically, picking customers from your client base who we predict will be your rolling 12-month’s best customers
4. Configure your audience.
Predict your customer’s future lifetime value from the first interaction. Nurture those showing signs of longevity.
Next, configure your seed. You have the option to compromise quality to increase quantity. For the sake of simplicity, let’s keep to the recommended amount determined by the Programmai algorithm.
5. Sync to your platform of choice.
We integrate with the leading ad platforms, including Facebook, TikTok and Snapchat.
Once you’ve built your lookalike audience, this will appear automatically in your Facebook ad account. We recommend running this predictive audience alongside other campaigns as an experiment into uplift.
So there you have it.
Real performance. Volume, efficiency & uplift. No compromises.
Powerful predictions built on your own identifiers, and a smarter way to serve your future best customers.
In the last few months, we’ve tested this dozens of times with our customers. Every time we do, we find that the performance is around double and the CPA around half of a traditional campaign.
For the marketer in a lab-coat out there, we’ve also Lift Tested this approach. And (spoiler alert), it’s always significantly more incremental. Win-win-win.
The future of advertising just got one step closer.