CXL - Ask them to spend more (yes, really)

Start dedicating more resources to encouraging customers to increase their spending. Don't miss out on a play that delivers exponential returns.

Keep reading for a great example of upselling and cross-selling in this issue. Plus, we're sharing how a marketing agency working with clients like Airbnb introduced self-directed learning in a rapidly growing team.

We also want to hear you – your calendar might already be tight, but if you block 20 minutes for a chat, you'll be part of the big changes coming to CXL.

TODAY'S HOST

Ričard Dechtiarenko, E-commerce Experimentation Lead @ Kilo Health

At Kilo Health, we work with a B2C ecommerce in the haircare industry. With a data-driven approach to our upsell strategy, we grew conversion rate 44.31%. Here's how:

The previous upsell sequence was poor: whenever users made a purchase, they received a special post-purchase discount for two other products of higher value.

No matter what they bought, the sequence remained the same.

Unsurprisingly, this ecom had only 8% of its revenue share coming from upsells. That's quite far from some of our other projects where upsells produce 20% of the revenue.

We optimized the upsell strategy with two key actions:

First one was analyzing what our best-performing upsell product was.

We tracked post-purchase upsell events in Analytics, looked at the impact of each product, and found the one with the highest conversion rate. That became our key upsell weapon and was offered whenever any purchase flow was completed. 

Second one was adding other personalized, more specific sequences depending on buyer behavior.

For one-time buyers, we offered a 50-60% discount to buy the same product. For subscription-based purchases, we instead provided relevant products related to what the user bought. 

We split-tested these approaches, with 50% of our buyers receiving the previously existing post-purchase upsell and the other half being exposed to the new upsell sequences. 

We ran the test for 14 days to achieve normal distribution and avoid any bias, and this was our minimal detectible effect for upsells, with a 95% confidence and an 80% power: 

After two weeks, the group which was optimized for our best-performing upsell product and with personalized sequences had a 44.3% increase in upsell conversion rate, with a 99% statistical significance.

AOV was also up 16%, and upsell revenue grew 40.79%.

The biggest problem with most upsell attempts is a lack of personalization. To be most effective, personalize your offers based on customer needs and usage.

Use data to uncover paint points, and highlight which premium products customers should know about.

Then, start testing and tweaking to ensure you're delivering the right message at the right time. 

Connect with Ričard on LinkedIn.

Cross-selling and upselling: what everyone is missing

Upselling and cross-selling are two ways to do the same thing: grow your revenue by getting customers to spend more.

But many businesses jump in too early, neglect their buyer’s intent, choose the wrong method, and annoy customers instead of enhancing their experience.

Let's discover what everyone's missing about cross-selling and upselling.

Read the entire article.

Customer Stories: Scaling an agency with self-directed learning

Here's how a marketing agency that works across 30 industries, wanting to integrate digital strategy and AI into their business, upskilled their team with cross-departmental knowledge.

CXL was part of the story too.

Read the Peak Ace story here.

Unpopular opinion: We can't all use AI

A few months into the AI era of SEO, we’ve already seen both great and terrible examples of placing AI deep into your content strategy.

Too many companies are trying to leverage ChatGPT and Bard to scale their content creation. Too many are doing it in the form of programmatic SEO, pushing out thousands of pages a month.

In some specific, rare scenarios, this is a growth lever. It already was before ChatGPT was a thing. 

Think Glassdoor salary pages, TripAdvisor ‘things to do’ lists, or some product comparison sites.

Problem is, it sounds like a silver bullet for content creation that’s highly scalable, reduces spend, and delivers great results. Not true, except on those occasions where there is zero competition for ultra-specific content.

If that’s your case, programmatic SEO is a safe bet as long as you don’t rely entirely on it. Otherwise, you might end up losing 80% of your traffic, as Gaetano DiNardi explained when calling it ‘problematic SEO.’ 

Here’s a very positive outcome of AI-powered content: it places more influence than ever before on those who don’t do it. You can only outcompete others with AI-generated content if they aren’t doing it too.

Original content, crafted by ‘organic writers,’ is going to be continuously differentiated from AI-powered strategies, as much as it is going to feed those AI tools. 

As Paul Graham puts it, we can’t all use AI. Someone has to generate that training data.

Connect with Pablo on LinkedIn.

How did we do?

Let us know by leaving your feedback:

First time here? Get these weekly by subscribing here.

Our address: 901 S Mo Pac Expy, Suite 150, Barton Oaks Plaza One, Austin TX 78746

Unsubscribe from future emails like these.