Webinar recording and transcript - discounting & promotions research findings from 2 Visions

Today, we will discuss discounting promotions research focusing on apparel. We worked with 2,200 US participants. This is a legitimate sample size, particularly for the US population.

There was a 2% margin of error at 95% confidence, which tells us this is enough to make some good decisions based on the data. If you want to go check it out, the report will tell you after the webinar at 2visions.org/promoreport; we've got some key findings there that you can check out. It's pretty long, and I'll break down some of those.

In addition, we also offer other vertical-specific research for free to the industry for eCommerce brands. This year, we've released one for home decor, one for personal care and beauty, and another for clothing and accessories. All of these have been focused on in-store versus online preferences.

Again, this is based on returns and warranties, Discovery channels where people discover products and loyalty factors, and we break those down per vertical. This year, we will also run with four, five, or six more. You can check those out at 2visions.org/ecommerce-market-research.

Discounts and promotions - key findings

Level of discount

Right off the bat, one thing we saw was when we were thinking about discount levels, there is an increasing need to increase the discount level we're offering. Many clients I've worked with have wanted to avoid this like the plague. They're afraid of playing an ever-increasing game with these discount levels.

And we're seeing that now. 20% was the minimum for favorable relative preference from consumers. It doesn't mean that ten or 15 may impact them, but it's not close to how it used to be.

For maybe a decade now, brands have used the 10% or 15% as an offer for email or text subscriptions. These levels have just been seen as throwaway numbers and no longer significantly impact the conversion rate.

So we saw that the 20% was a psychological level on the floor now. 50% plus moved the likelihood of buying to 99% if the product was loved. If the product is loved by the person looking at it more than liked, and they view it as a product they have to have, if the discount is 50% or higher, the likelihood to purchase went up to 99%.

Uniqueness and scarcity

We studied the combination of discounting alongside affinity, which was how much they liked or loved the product, but also aside from the uniqueness of the product.

If they perceive the product as being highly unique, and for clothing, this is important. You know, denim is denim, a shirt's a shirt. But if they view it as exceptional, maybe there's a texture or function to the item. How much did that play into purchase decisions?

The other factor was scarcity. If they knew that this item would always be available, did that impact the purchase decision versus if they thought that it would be gone and not be available in the next four weeks?

We found that uniqueness and scarcity when measured against the discount level and the product's affinity, while influential, were negligible compared to the other two. These worked best as items to bundle and make as closers to tip someone over an edge.

But it wasn't as influential as the discount and the product's affinity. Moving from liking to loving an item was paramount. It had a similar impact on purchase behavior as moving from a 25% to a 50% discount, which is a sign. I mean, it's a significant move there. Liking an item was relatively neutral. They kind of just assumed they were going to pick items. They didn't look at choosing an item as really a great thing. It was very neutral.

You can't avoid the focus on the actual product fit and positioning, and you won't be able to discount your way out of that. It's important to note that if you're looking to study something to make the best use of your discounts, you want to study how much the customers love what they're buying and are not on the fence about it.

We also examined whether they felt unsure about something; if they did, it was a solid negative relative preference.

Discount fever is here to stay

I wish I could say, hey, there's no taste for discounts anymore, so we can give up on that. But that's not the case.

62% said they primarily delayed clothing purchases until they were discounted. That doesn't mean they won't find themselves buying something that's not discounted, but it does mean that the majority specifically have a strategy of waiting to buy until it's discounted.

This means that discounts must be significant in how we merchandise and market. We must discount well if we want to significantly impact conversion rates and customer lifetime value.

The competitive advantage of discount engines

Software systems that allow you to target and tailor these promotions are essential because you can only dedicate so much money to discounts from the top line. It would be best if you used that money very well so that it changes the business.

For many clients I work with, our average annual eCommerce revenue is about $500 million. We have some huge clients and some small ones and everything in between.

But we don't see much intelligence in using these discounting engines; they lean on the engine to do most of the work. If you can think about an internal take on how we tier tailor or personalize discounts, that can become a competitive advantage for you.

There's a lot to think about daily. There are a million moving pieces, and this is just one more of them. But if you can get away, light a candle, put on some music, take 20 minutes, and start to jot down, okay? If I could take our discounting budget and spread this in a way that could activate, how would I do that?

And using third-party software systems is going to be crucial.

The importance of affinity

When a shopper is on the fence about an item, only the 50% plus discount level makes any difference. You couldn't move them with a discount if they didn't like the items.

You could throw 40% out there; they're going, "Maybe not."  We ran a test - a group of people who loved a product, and we gave them 10% off versus people who were unsure about a product, but we gave them 50% off.

When we ran that simulation on the model of the entire us population, the market was a 50/50 split. Half of the people would buy it at that 50% plus level, but the other half would say no, and they'd go for the one they loved at just 10% off.

eCommerce brand takeaways

Discounting well is possible, but it takes more targeting intel and testing effort than brands dedicate

How we discount and learn from our discounting behaviors is essential for considering discounting as a hypothesis.

Just like we do with UX—A/B testing or any sort of performance testing, we want to have a hypothesis, and then we need to activate it through systems. We then need to learn from that. Much of that learning can be programmatized, and the system can do it.

But you want the humans inside your organization to know what we're learning about, who you can sell to today, how they're responding, and how the different segments of your customers are responding to these discounts.

Let the program do some of it. But you also need to know what made a significant impact when we changed. How we discounted the offer medium matters.

In the research report, we tested Black Friday, Cyber Monday coupon codes, and email subscriptions. When we looked at these different delivery methods, we also saw site-wide friends and family, flash sales, and these sorts of things.

All of these offer mediums matter. It's a targeted test. There isn't just a silver bullet.

"Hey, this is what's going to work."

It would be best if you tried for your brand. And the most important thing is that you can learn rapidly what worked and what didn't, and then you can change and test repeatedly. If you do that, you will be at the top of your competitors. Not many people are doing this very well right now.

The offer level matters

What we're seeing is that different segments of your customer database will react to these levels differently.

When we looked at the cluster, we noticed that millennials in Gen Z really cluster in preference around these levels, and baby boomers and Gen X cluster more similarly in their preferences. We're also seeing differences between men and women in offer levels that are more similar in their preferences—where and how they prefer.

Consider when you might throw out more drastic offer levels and how you'll do it so that maybe it's not site-wide, or you can tier it with different cart sizes. Still, you want to have a hypothesis around it and throw out a goal of what sort of CVR increase this will make for you.

Let's say you offer a 40% discount. Assume this will increase the conversion rate by 20% and throw something out there. Whether you will hit that or not, we'll learn about it.

When you document this, not just the level, but also what you throw it on, how you get it out to them, and the medium that offered the discount.

You cannot afford not to refine your product offering competitively

Many companies love the smell of their roses. You can see people when they're walking through their product; they go, "These are great, people love these."

But competitively, they may need people to love them more. And it's hard to find out.

How do we do that?

You can also research, just like we did on the market. We work with companies that run research on the product themselves, whether this is color variance or functionality, the naming or the features, or whether it's wicking or breathable.

All of this can be researched and tested so that you develop what we call more tier-one products—hero products. These sell well because the more of those products you have, the better your discounts will work in the long term.

You're not just moving products that people kind of like. Market-level intent is directional. What we've provided in these research studies is directional data. The fact that it is vertically specific helps it to be even more directional.

You can narrow this down to your specific brand niche, to you and a few competitors, and that's actionable data. Take the stuff you read in a market research study like this with a grain of salt and use it to form hypotheses.

But the kind of research you can run for your brand allows you to have more nuanced, concrete variables that we can test. The data model fit improves, and the relative value changes among products and brands. You get very actionable data.

Q&A

BFCM (Black Friday, Cyber Monday), we sometimes call it peak, whatever you want. At the end of the year, that period is where a lot of eCommerce performance comes from. Did you learn any specific things about that from your research?

Yes. It's so interesting. Every time I study with research, I learn something new. One thing that we've known, by the way, is that senior executives often come from the field. They were heads of sales or marketing. They had a closer touch to what was going on in the market.

Time has passed. Five years are one year, and one year is six months. This means we must run this research more often because customer preference patterns change.

For BFCM, I considered going into this, and it's played out. People have been doubting if they're excellent deals anymore. I don't know if you noticed over the last year, there was way more traffic on social media. It felt like people were questioning, "Is this a scam? Are these deals even excellent?"

I remember seeing posts of companies that had the product discounted a couple of weeks earlier for a lot more. Then, we saw in the research that there is still a high preference for Black Friday and Cyber Monday; it was 58%.

People thought that the best deals of the year still came there. More than the email subscription, more than the coupon codes, more than anything else.

They thought, "This is how I will save the most this year."

It's still relevant, and we still need to play the game. I've worked with companies that have tried to avoid playing that game, and they said, "Look, this is too much. We don't want to jump into the hoard and get buried. We're going to play it safe."

Last year, we saw a lot of lost sales volume. Instead of not playing, the best thing to do would be to play it smarter, start dialing in your strategy for BFCM, and test again.

The goal is to use this period to increase customer lifetime value. That's a much better way to justify the increased discounts that you're offering. So yes, you're offering them, but challenge yourself to make it have a higher impact than what you're used to seeing when you roll out a discount on the site.

We also saw that males were more sensitive and influenced by BFCM than females, which was 15% higher than females. I would still consider BFCM part of your strategic discount timing during the year.

If you have the right size with tier-level discounts, 20% off for $100, you know, whatever it is, you go up. Plus, if you offer a stackable code, if you play your tiers, assuming a stackable code, that would be the ultimate volume play. People love the game of all of this and feel like they're getting over on the company.

That's what BFCM is all about. They have to believe there's a reason for this. Play into that story.

Some people prefer newer brands, while others prefer older legacy brands. Are there any differences between the people who like new or legacy brands?

Yes. When we tested, we didn't know what to find, but we saw some differences.

Those who preferred the newer brands had a higher preference for highly unique items. You could guess that when you think about it, but their preference for more highly unique items is there.

If you're a new brand coming out, know you have an advantage over the incumbents. If your item is perceived as unique, does it have to be unique, and to what degree? You have to research that and test it out.

You'll have an advantage if consumers perceive it as highly unique and love new brands. They also had a higher dispreference toward items perceived as typical.

They were more forgiving on product affinity, which surprised me at first. They were almost more willing to jump onto the exciting new brand game and were more forgiving if the product had to be awesome because they just wanted to be part of this new thing versus those who were more for the tried-and-true legacy brand. They had a much higher need for the product to be loved.

The 25% to 30% discount levels move the needle more for those who prefer new brands. If you're a new brand and you're throwing out a 20% discount, you're not going to get as much bang for your buck in terms of movement in sales volume as you would at the 25% or 30% level if you are a new up and coming brand.

What are some of the key differences you found between those different generations?

There wasn't a ton across the whole thing, but some key areas had significant differences.

We saw clusters. Sometimes, you don't see this. You don't always see the millennials in Gen Z, Gen X, and baby boomers cluster into older and younger groups.

In this case, it was robust clustering. For instance, coupon code trust was 50% higher when you were the Gen Z millennial versus the Gen X boomer. This may be because they've had more experience with it.

If you think about coupon codes, in many cases, you've got to know to gearch the Internet for them, you have to have an add on in your browser where the coupon codes come up, that sort of thing. It lends itself more to a digitally native generation who knows that the option exists.

It's the same thing wiser with prescription; millennials value brand subscriptions the most moistened ast. So again, with those subscriptions, we saw the younger folks more likely to believe, hey, I'm going to get a deal from this thing.

Gen X and Boomers are not even really looking for it as much.

We get asked often, "What is the right level of discount? Where should I be pitching things?"

This research was predictive. It was done with predictive analytics. This wasn't just a survey. studied

their behavior around purchase decisions in a hat they didn't even know they were doing. And then, we calculated the model of us consumers, and how they would act. I could also play with this model and put different saddlations into it.

We found that a 20% discount doubles the likelihood of purchasing. The 50% we mentioned earlier leads to a 99% increase in the likelihood of buying.

The 20% was the psychological threshold that we saw. We're studying these thresholds not only for discount levels but also for pricing levels. If I price this at $9.99 versus $12.99, is that a threshold where I see greater or more limited product adoption?

You can find these levels out.

We saw that 20% was a big deal and could change next year.

We also saw that women were less swayed by 50% off than men. Even if they just kind of liked it, they're like, let's do this. The women were more careful.

What is more effective, scarcity or discounting?

Discounting is far more likely to influence purchase than scarcity. I thought maybe we would see it be a little closer.

Scarcity and uniqueness were not significant factors, except when used as a tipping point, right over the edge if someone considered it. That doesn't mean it's not influential. It just means if you're asking about, well, is it influential relative to discounting? The answer is no.

The discounting: You could say this will be in stock forever, but today, it's 35% off. That will be much more impactful than saying this sweater won't be around in three days with scarcity and a discount. If you combine them, they are 178% more likely to buy. It's a great combo when you put these things together.

Discounts were more direct; they had a more substantial impact, but scarcity boosts value, and it was best bundled with the discounts.

What lengths do consumers go to to get discounts? Will they go quite far to get themselves a good deal?

They absolutely will. And it rhows you we should be planning to give them ways.

If 62% of customers delay purchases to receive a discount, instead of fighting it, we should really find a way to embrace this and go; how do we give them ways to do what they want to do?

They want to know a specific type of buyer, not every type of buyer or shopper, but there are many that they want to win. They want to know they've been responsible for their money, particularly in this economic climate.

They want to know, hey, I didn't waste my money. One way that they know that is if they get a discount. That doesn't seem normal if they feel like they've won, particularly if they can tell a spouse they've won.

"I wasn't going to get it, honey."

My wife would come home with things and tell me how much they were on sale. She wouldn't tell me how much she spent, but it's the same psychology type, so it matters. They will use apps; they will use tools for this.

Remember, sometimes we just try to offload products that aren't winners, and we use discounts to do that. That's okay. But I don't want to lose sight of the fact that there's a real potential for discounting to be used on highly winning products, to take scale to a hockey stick level, to really take it to the next level. And there's something that you can do in volume that really justifies the discount, particularly when you see the changes to customer lifetime value.

If you're not calculating that right now, then you don't know what the impact these discounts will have to lifetime value.

It will be a lot harder to justify it, and then the people who can explain it can offer the discounts, and they will get the scale. It behooves you to begin to track that and understand the relationships.

It doesn't look like a discount would persuade someone who didn't want something to buy it

If people perceive the product as cheap, again, 50% off or higher, it would influence them if they were on the fence about it.

Anything less than that was less likely to influence them to purchase.

If they don't like it, they won't buy it.