How to stay in the 0.5% chargebacks “safe zone” for digital goods
Fact: industries such as eGifts and eGaming are under constant threat of fraud.
According to a recent study sponsored by PayPal, the primary challenge online businesses are facing is battling the increasing sophistication of fraudsters.
This is followed closely by not having the right tools or practices in place to mitigate online fraud, which doesn’t make life easier.
This reality affects your bottom line.
How are online merchants losing money, exactly?
As an online merchant, you’re probably living in a reality where you still have to manually review transactions as your last line of defense.
Of course, employing a team of fraud detection experts may be effective to a point, but they are expensive and slow.
This generates a bad customer experience due to delayed delivery (it takes time to review all the relevant data - anywhere between 30 and 120 minutes per purchase).
For high-volume sales environments where immediate fulfillment is key, you should avoid this by all means.
Another option is enforcing strong rules and/or restrictions for all customers. This basically means blocking out any suspicious-looking transactions.
This naturally adds some type of friction to the buying experience and sometimes can reject your legitimate buyers, leading to a significant drop in transaction conversion rates.
Look, we understand. Times are tough, but I am happy to say there is a cloud of hope amid all this negativity.
In this post, I’ll show you a better way to not only combat fraud but prevail against it and stay in the 0.5% chargeback “safe zone” for digital goods.
Let’s first make sure we are on the same page:
What is the chargeback “safe zone” and why 0.5%?
A chargeback safe zone is the acceptable percentage of transactions that are labeled as fraudulent within the payment processing cycle.
A higher chargeback ratio determines the risk factor and ability to process payments.
It’s widely accepted that merchants with a consistent fraud rate of 0.7% of all transactions are considered in the “safe zone”. The zone between 0.7% and 0.9% is often referred to as the ‘danger zone’ where businesses have to be extremely careful not to end up, as it basically means you’re flying too close to the 0.9%+ sun.
Those that end up with a fraud rate greater than 0.9% will be outright flagged as high-fraud merchants. This means two things:
- Incurring major fines from payment networks for every transaction labeled as fraud.
- Getting blacklisted from accepting online payments altogether, which is a fast lane to potentially shutting down the business.
Now, the question I’m pretty sure is on your mind: why 0.5% and not 0.7%?
The 0.7% rule of thumb is a designated safe zone from the payment network's perspective.
The thing is that the margins are very small. And so, every transaction counts and it’s very easy to “wander” into the danger zone and get slapped with fines.
We believe that
0.5% and below should be the upper limit for every merchant in order to avoid any kind of fines or worse, risk being blacklisted.
I cannot stress enough that as a merchant, you are solely responsible for your individual fraud/chargeback rates.
Here is how fraud prevention impacts your purchase conversion rates
The pandemic-driven shift to online transactions was always going to put front and center one key challenge:
to beef up security without causing too much friction for new and loyal customers.
While optimistic in their intent to successfully solve this challenge, many digital goods businesses fall short.
Based on our extensive experience in the fraud protection field and independent research, we found out that the fraud management framework set in place has a ripple effect across the general conversion rate of purchases.
By safeguarding against fraud, you actually create friction in the pre-purchase process, in the form of:
- Blocking certain geos
- Maximum purchases per day per IP address
- Maximum payment allowed per card per 24 hours
- Payments with 3-D Secure/PSD2 implemented
- Two-step verification
- And more
However, these friction points routinely lead to false positives/declines.
You indeed get to keep away a significant amount of fraudsters and fraud schemes from harming you - that is, until they figure out the "rule" and then circumvent it.
At the same time, you also push away your legitimate customers due to overly tightened acceptance parameters.
You get it. You end up with low conversion rates and a worrying number of false declines at the point of purchase.
It’s not just about false declines.
It’s also about your customers-to-be who are unwillingly directed to another seller, meaning that you lost revenue from that individual purchase AND future ones from a potentially loyal customer.
About 40% of new shoppers simply won’t come back for another purchase after experiencing their first decline. These lost customer relationships are never recovered.
When you add the long-term damage caused by lost loyalty, friction-driven failed sales, the actual cost of digital goods fraud becomes far bigger than expected.
In our experience, removing these friction points can increase the conversion rates by more than 100%.
This is because the entire purchasing process becomes more enjoyable (higher conversion rates), which leads to continued sales to the same buyer (brand loyalty).
In turn, the good word spreads across social media (positive brand image), creating a neat loop.
We all know it’s typically far easier and cheaper to get repeat business from existing customers than it is to win new ones from scratch.
For all of these reasons (but also because it’s common sense), it is critical to consider the true influence of a fraud prevention system on purchase conversion rates.
How to remain in the “safe zone”
I have three words for you: Predictive Artificial Intelligence (AI).
AI is the reason why we have liability shifting today - the best fraud prevention model the cybersecurity industry has to offer.
Liability shifting is the higher security standard in which fraud prevention companies assume their clients’ fraud liabilities.
This represents a significant change in the handling of payment fraud as the goal isn’t just to aid businesses deal with the risk of fraud but to completely eliminate it from their ranks.
Thus, liability shifting has a highly advanced technological component as it applies a machine learning algorithm that learns autonomously from mountains of unknown data.
Thanks to continuous training and learning, it can differentiate between fraudsters and genuine buyers.
In fact, it is also able to tell apart nuances of actual yet unusual purchases. A confused grandfather getting a little something for his grandchildren or a purchasing agent buying supplies over a legitimate VPN will both rightly account as legitimate transactions.
AI has the ability to make these assessments in a blink of an eye and with as little customer disturbance as possible. It allows you to provide a friction-free customer experience and still keep up with whatever fraudsters are scheming.
Unlike physical goods, the immediate delivery of digital goods requires real-time, automated decision-making to approve or decline the transaction.
I’ll go even further and say that in today’s high-tech world, manual detection should be viewed as a weakness, even though it acts as the last line of defense in many cases.
Human fraud detection teams are expensive by nature, as they are slow and not completely resistant to errors.
And the numbers speak for themselves.
Research suggests that the annual cost of manual reviews for a small merchant is roughly $378,000, while the figures rise up to around $825,000 for medium-size merchants.
If your current fraud protection setup isn’t continuously monitoring for fraud and/or is overly reliant on rules and restrictions, you are likely rejecting legitimate customers.
So, the solution for a comfortable spot in the chargeback safe zone is a predictive AI fraud protection system that makes accurate suggestions and decisions in real-time by finding the optimal balance between a healthy chargeback rate (under the 0.5% threshold), no rules, and low manual review costs.
Not all AI is created equal
Singing praises about AI is one thing I’ll be more than glad to do any time of the day.
Still, I wouldn’t be much of a credible person if I failed to mention that AI will likely never be perfect.
Despite being far and wide better at analyzing online purchase behavior more accurately and at scale than any human team, AI/ML models still aren’t 100% right.
Our extensive industry research found that the majority of AI/machine learning models can accurately approve only 85% of purchase attempts.
These represent the clearly legitimate buyers, which is about 84% on average, while the remaining 1% falls on fraudsters.
Then, there’s the issue of how most solution providers implement liability shifting.
The remaining 15% of the purchases are being rejected in order to be on the safe side.
I stand by my words that liability shifting is the state of the art in fraud prevention but there is no denying there is room for improvement in terms of its accuracy.
Another important thing to note is the generality of existing fraud prevention solutions.
Whether we are talking about crude identity management to advanced machine learning tech, the majority of options on the market were developed with e-commerce sales specifically in mind.
Where does that leave your business in dealing with digital goods fraud?
Short of the finishing line, I’m afraid.
To be fair, these products do their job for the most part and minimize exposure to fraud risk, but they also inadvertently do some harm as they treat all online sales alike.
This is a time of continuous change where each product category has its own challenges to overcome when fraud comes knocking on the door.
And what’s happening now in digital goods is no exception so each category calls for a tailored solution for maximum results.
The 85% mark is what a standard AI/machine learning model is able to correctly determine when it comes to the legitimacy of payment purchases.
I say ‘standard’ because we at nSure.ai have raised the stakes… We’re proud to guarantee a 98% transaction approval rate for digital goods.
Here’s how you can have peace of mind and sell digital goods with confidence.
What makes the best fraud protection solution
The first step to the best possible protection for your business is to get a holistic understanding of how each piece of the fraud-fighting puzzle fits in the big picture.
In that regard, there are three key elements that ultimately make the best fraud protection system:
- Hundreds of different data points for analysis
- Segment specialization
- High level of transparency
For instance, nSure.ai examines more than 500 data points and their combinations in its relentless hunt for fraudsters.
A large part of this data isn’t even understandable to humans so running it through our algorithm offers a higher level of specificity than any manual review.
Analysis of various data points is usually a mix of three loosely categorized groups: Contextual data, Behavioral data, and Account data.
Contextual data refers to “passive” features of the transaction, such as:
- IP address of the potential buyer
- purchase size
- device type used
- their browsers
- and more.
Behavioral data concerns the “actions” of the customer, including:
- how much time they spend on the page
- the time elapsed between entering the website and attempting checkout
- whether they typed or pasted the password
- opted for a discount or paid the full price
- and more.
Account data refers to all the details tied to the account making the purchase. These include:
- the age of the email used to create the account
- verification of the phone number
- address of both the cardholder and the shipping address (if applicable)
- and more.
In some cases with a comprehensive fraud prevention system, the services of third-party data vendors are retained. This allows the analysis of outlying data points that might be relevant to the overall decision of which purchases to accept or decline.
Data is critical for any AI fraud-fighting effort as a machine learning algorithm is only as strong as the data it is fed with during training.
That is how we are able to deliver over 98% accurate approvals, declining only 2% of your sales, compared to the 15% industry average.
In that spirit, segment specialization allows the focus to be primarily on specific product categories across the e-commerce landscape, instead of treating all the transactions the same.
Speaking for our segment, the challenges of digital goods are unique and have to be handled as such.
Products like digital gift cards are ideal for fraud operations at scale as they personify speed and convenience. They are sold and delivered digitally and immediately, are anonymous, and can be easily resold.
Now compare that to a merchant selling physical goods online. It’s hardly the same, right? There is less data to work with, which requires better use of the existing information.
This is why segment expertise.
And you should always demand a certain level of transparency from your fraud protection vendor.
Still, making decisions regarding which transactions are genuine and which aren’t in real time means it’s somewhat difficult to exactly point to the factor(s) for each declined transaction.
But knowing the reasoning behind these decisions allows you to keep your finger on the pulse of your approval and decline rates. As a bonus, we offer a 100% chargeback guarantee, providing yet another layer of confidence in what you’re doing.
”For the times they are a-changin'”
In the words of the immortal Bob Dylan, the times they are a-changin’ and fraud is very much included.
There is a distinct lack of regulatory oversight and industry standards for digital goods, which mean two things:
- There is very little protecting companies of all shapes and sizes from chargebacks caused by these fraudulent transactions, as well as other types of digital goods fraud.
- Fraudsters can freely come up with effective schemes that target you, an honest merchant going about your own business - literally.
So, it’s not surprising to learn that fraudulent digital transaction attempts against businesses increased 46% worldwide and 22% in the U.S. between March 2020 and March 2021.
These attacks aren’t going away. If anything, they are likely to continue at a steady rate, if not incrementally increase.
Successful fraud prevention relies on sophisticated innovation driven by predictive AI.
But it also relies on you.
On your intelligence, experience, and instinct to focus on the right practices and policies that will help you steer clear from chargebacks and rejecting legitimate customers in the first place.