How to Increase Sales Using Machine Learning For Cross- and Upsell?

Photo of Grzegorz Mrukwa

Grzegorz Mrukwa

Updated Feb 14, 2023 • 8 min read
phad-pichetbovornkul-269226-unsplash

The final success of each and every business or project comes down to selling.

It has always been like this, but nowadays - when we’re able to get ahold of massive amounts of data about our customers - it is just easier to focus on building relationships with them, and then make the most of this in terms of sales.

An approach to managing relationships with both present ando potential customers - also known as customer-relationship management - is something that is absolutely essential, if only because it affects a company’s revenue. It also builds loyalty and improves customer satisfaction. How? In a nutshell, by predicting customer needs and providing them with tailored offers right on time.

These days, having customer data close at hand, you don’t have to navigate blindly. You can use the information you have and take advantage of it to better interact with your customers. And by better, I mean more effectively for both sides. Plus, you don’t necessarily have to use fancy CRM systems (however helpful they may be) to streamline the communication process. You can basically start with any data you have right now, from demographics and purchase history, to the frequency and style in which your customers use your app.

Having information like this, you can start to work on implementing some of the most popular sales-improving tactics: cross-selling and upselling. What are they all about? Check out the details below!

What is cross-selling and upselling?

Generally speaking, cross-selling is when you suggest related products that go outstandingly well together with your customer’s original purchase. For example, when someone chooses a pack of coffee, they might also be interested in buying milk.

Upselling is slightly different. This is about persuading a customer to buy an upgraded version of their desired product - one which is better and obviously more expensive. For instance, if a customer wants to buy a basic coffee machine, you may suggest that they consider a fancier appliance - one that features a built-in coffee grinder and a timer.

So, which one of these tactics should you incorporate in your sales strategy? The answer is: both. Just make sure to embrace some solid, data-driven techniques.

Analytics is king

First and foremost, each and every suggestion that you give your customer has to be relevant. So, what you need to do is focus on profound data analysis before you start developing a specific action plan. Having the best possible understanding of your clients, you will be able to interact with them in a much more effective manner.


There are two things you can do at the very beginning:

  • Customer profiling and segmentation. The first concept is a way to create a portrait of your typical or ideal customers. The latter is about breaking them down into groups that share the same demographic (like gender, location or age) and psychographic (like values, hobbies or habits) traits.

  • Market basket analysis. This is a technique developed to understand customer purchase behavior by analyzing what products are often put in the same basket and bought within one order.

Having data like this and knowing what your customers value, you can create recommendation systems from scratch or just polish your current systems to address your cross-sell and upsell campaigns. You will be able to decide which products should be presented together as a perfect duo, or make improvements to the presentation of search results in order to influence the customer’s subconscious. You may even want to go a bit further and make some hot and highly personalized offers, using sophisticated algorithms and advanced predictive models that were created to analyze transaction data and trends. For example, according to a McKinsey report, 35% of what people buy on Amazon and 75% of what they watch on Netflix already comes from recommendations based on such algorithms.

Putting it simply, data gives you the chance to develop some serious strategies. Speaking of which...

Strategies, tests & challenges

There are tons of different strategies for both cross-sell and upsell campaigns - just based on the data you gather from using the above-mentioned techniques. You can either try one strategy at a time or loosely combine a few different techniques - it depends on the type of your business and the goals that you have.

Cross-selling and upselling strategies

When it comes to frequently used cross-selling techniques, most of the time this refers to you recommending products to your customers that are:

  • supplementary (like a lens to a camera),

  • or related (like weightlifting shoes and workout shirts).

Notice that you don’t even have to make use of any purchase insights to come up with recommendations like these! They are just logical.

You can also offer items that other customers who looked at similar items tend to purchase, fancy gift wrapping or some massively discounted products. Such incentives can make the overall value of the customer’s basket go up significantly, along with their satisfaction.

Upselling, on the other hand, is much more complex and usually requires more personal data. For example, you could offer an upgraded subscription plan for a mobile phone to a woman who regularly exceeds her Internet data usage, or a better computer to a man who has already looked at a number of different gaming laptops in a store. Upselling strategies also include:

  • featuring new arrivals, time-limited offers (like seasonal or daily deals), and bestsellers,

  • suggesting products that have been recommended by other customers,

  • offering free shipping above a predefined order value to encourage people to buy more expensive products.

Both upselling and cross-selling techniques can be used:

  • before the purchase (i.e., by featuring products on the homepage),

  • during the purchase process (i.e., by offering additional options alongside a product),

  • after the purchase has been completed (i.e., by recommending trending products in an email receipt).

And one more thing. All of these campaigns should always be verified to see what performs really well, what does not produce any results at all, and what requires only a few minor modifications. You can measure the conversion rate for various upselling/cross-selling strategies with a series of A/B testing, and adjust them accordingly. But, seeing as it’s almost 2019, it may be smarter to do this together with...

...a little help from machine learning. Why?

Because one of the biggest challenges for businesses nowadays is incorporating analytical insights into products and real-time services to make customer targeting much more accurate. Working with ML-based systems can be a game-changer, helping you make the most of your upsell and cross-sell campaigns. And we know exactly how to do this, as building highly efficient and innovative software is our bread and butter.

ML-powered sales campaigns can help you simultaneously increase customer satisfaction and brand loyalty, affecting your revenue remarkably. This is an investment that every company will have to make, sooner or later, in order to maintain their competitive edge. So, if you are still on the fence about this - don’t be. Just get in touch with us, and let’s talk about your options.

Photo of Grzegorz Mrukwa

More posts by this author

Grzegorz Mrukwa

Former Data Science Manager at Netguru
Lost with AI?  Get the most important news weekly, straight to your inbox, curated by our CEO  Subscribe to AI'm Informed

We're Netguru

At Netguru we specialize in designing, building, shipping and scaling beautiful, usable products with blazing-fast efficiency.

Let's talk business