Quel pourcentage de votre désabonnement devrait être dû à des échecs de paiement? Que pouvez-vous faire pour diminuer ce pourcentage?

Usman Gul

Usman Gul, Croissance @Marqeta. Anciennement @MasterCard

Répondu il y a 181w · Voté par

Sankaet Pathak, Fondateur @SynapsePay

Payment failures are an epidemic in eCommerce. The architecture of the credit card processing tunnel was designed for the pre-Internet era. The system was not built for CNP, and doesn't work well for CNP transactions.


The Quick Answer:
Ways to lower transaction decline rates range from building a best-in-class in-house payments team to hiring consulting firms that specialize in reducing decline rates. You can make modifications to your transaction routing system and to your payments gateway that can reduce your decline rates. If you have the means to hire a consulting firm, or to build an in-house payments team, then you should do it. It will pay off.

Also, an effective way to lower your decline rates (and increase your conversion rates) is by accepting payments via direct debit or ACH-based payments systems: Plaid, Dwolla, or Synapse Payments.

Now, the details...


Its An Ugly Situation:

  • 15-20% of all card transactions in eCommerc are declined
  • With recurring charges, 32% of total transactions across all industries are declined
  • A vast majority of all declined transactions are unrelated to security/fraud
  • A majority of fraud or security related declines are legitimate transactions that should not have been flagged as fraud ("false positives")
  • Per PayPros research, for small businesses with average annual revenues of $110K, declined transactions result in a loss of: 4.97 customers each month and $612 in monthly revenues. These businesses also incur $271/month in soft costs (communications, administrative, etc.) due to declined transactions.

The problem is much bigger for large businesses.

Source: Publicly available MasterCard, Visa and PayPros research.


Why Legitimate Transactions Get Declined:

Here is a quick overview:

  • (I) Invalid Card Numbers:

Cards Expire: Debit/credit cards have a short card life (<2.2 years). Prepaid cards have an exceptionally short life (<6 months)
Lost/Stolen: Cards get lost all the time. Cards also get stolen

  • (II) Technology Failures:

Communication Failures:
Response Code 63: "Security Violation"
Response Code 30: "Format Error"

The architecture of the debit/credit card tunnel requires each transaction to go through at least six different entities (processors, banks, networks, etc.). Each entity must be able to flawlessly connect and communicate with all others in a common language, and each signal sent must meet the strict (and unique) security requirements of the receiving entity. As you can imagine, that is not always the case, especially when a transaction is initiated via (and returned to) an online payment gateway. 'Timeouts' are an extremely common industrywide problem that is caused by a signal that does not fully meet the security requirements of the receiving entity, and thus the receiving entity is unable to respond to the signal within the allowed time. Here is a diagram that I use to help new hires understand the wiring of our payments ecosystem:

Quel pourcentage de votre désabonnement devrait être dû à des échecs de paiement? Que pouvez-vous faire pour diminuer ce pourcentage?

Mapping Failures:
Response Code 80: "Network Error"
Response Code 96: "System Error"

The below diagram illustrates how each credit card is linked to a BIN, and each BIN to an ICA. This structure requires massive mapping tables that can map each BIN to the right issuer, acquirer, processor, network, etc. The process to enter data in these tables is mostly manual and hence prone to errors. Its also not easy to change this system (because all acquirers, all issuers, all issuing processors, all acquiring processors, and all networks use this system). To make any foundational change to this system, you need buy-in from 100+ really big, conservative institutions. Not easy.

Leggi:  What are the legitimate ways to earn money online?

Quel pourcentage de votre désabonnement devrait être dû à des échecs de paiement? Que pouvez-vous faire pour diminuer ce pourcentage?

  • (III) Security Failures

False Positives:
Response Code 05: "Do not Honor"
Response Code 57: "Trx not Permitted"

Issuing banks use a risk-based authentication (RBA) system that relies on probabilities. The system assigns a risk score to each transaction. Purchases that do not conform with your general spend behavior are assigned a high risk score, and often get flagged as fraud. As part of this process, many legitimate transactions are incorrectly flagged as fraud ("false positives").

The issuer-based RBA system turns out to be a terrible solution for online transactions. In eCommerce, the way to determine fraud is to use advanced techniques such as on-site click behavior, keystroke analysis, optimized credential management, etc. Issuers can't do that because they don't have web navigation data.


Bottomline:
You can minimize transaction decline rates to less than 10% if you have the capacity to build a cross-functional, in-house payments team. Or, you can hire a consulting firm to do the work. Either way, it'll be expensive and will probably not be worth it for businesses with annual revenues of <$50MM.For business with <$50MM in online revenues, or for teams that don't have the resource or time capacity to resolve this, I would recommend looking into direct debit or ACH-based payments systems: Plaid, Dwolla, Synapse Payments, etc.My personal favorite is Dwolla. Its a full service Internet-based payments network that eliminates a number of problems:
  • No concept of lost/stolen or expired cards - Dwolla accounts don't expire and cannot be lost
  • No technology failures - a transaction via Dwolla doesn't involve 6+ entities, and doesn't rely on an archaic BIN/ICA mapping table
  • Fewer false positives - they own the web platform where a payment is initiated, so they have access to data that RBA systems at issuing banks do not. They definitely have fewer false positives, although its not clear what the delta is.

Of course, credit cards have their own set of advantages but payment failures and churn rates is definitely a huge leakage point w/ card payments.

Disclosure: All views are my own.

Floyd Earl Smith

Floyd Earl Smith, NGINX writer

Répondu il y a 190w

The estimates we use at Recurly are 9% card failure for B2B and 14% for B2C each time you run cards. That's triple to quadruple the 1/36th (roughly 3%) you'd expect due to cards dating out after three years. Those initial card failures end up as varying amounts of churn.

Having a charge fail is a big hassle for you, and customer contact around this is a time-consuming for both you and the customer, and embarrassing and otherwise not positive for the customer. So it's better to avoid the failures if you can.

At Recurly, this is a big part of what our subscription billing software does. We cut initial charge failures, intelligently retry failed charges, and do automated dunning on your schedule and with your content. As you might expect,this recovers revenue, reduces payment-failure-related churn, and produces a better customer experience. We describe a lot of this on our Blog | Recurly and elsewhere on our website.

Leggi:  How to find skilled ICO developers for my ICO development project

Alain Mevellec

Alain Mevellec, Co-founder at Sellsy

Répondu il y a 237w · L'auteur dispose de réponses 1.9k et de vues de réponses 2.1m

If your product is good, this should be your main churn provider.

If your product is really good, this shouldn't matter: customers will just update their info.

Anyway, credit card will never be a sustainable source of income: expiry date, stolen or lost cards... You can't really rely on those.

Rules of thumb:

  • offer long term formulas (6 months/1 year) and let people cumulate them if they want (as an example, at Sellsy, we have people committed until 2017 ��
  • offer others sources for payment (wire/check/paypal)
  • if you get bigger customers (with higher ARPU), offer them long time contracts with monthly debit (you can add a little SLA or free training/setup to justify that).
  • another thing we made at Sellsy: we always bill options/seats for the remaining part of the subscription. Example: if you commit for 12 months and add a staff after 4 months, you'll pay 8 months for this staff. This is easy extra cash, as people are already enjoying the software and don't really care.

Christine Speedy

Christine Speedy, Passerelle de paiement B2B et fournisseur de solutions SaaS compatibles PCI pour les cartes absentes, mobiles et le commerce électronique. 3D ...

Mise à jour il y a 141w · L'auteur dispose de réponses 246 et de vues de réponses 272.5k

2016 updated DUE TO INDUSTRY CHANGES: There's no 'suggested' churn rate. It can vary by industry, product or service sold, or industry changes like EMV that create abnormally high replacement of cards in the market.

Regarding the second part, try a multi-faceted approach. Here's some options:

  1. Visa / MasterCard card updater service. The acquirer, gateway, and merchant account must all support the service. Merchants pay a one time application fee for each card brand and then a small fee per card updated. Merchants apply for the service through their acquirer. (If merchant changes acquirer, then must reapply for service again.)
  2. Automatically deliver notices (and repeat if no response) in advance of card expiration with link to update; some gateways do this, some don't.
  3. Reduce friction to update- Empower customers to self update payment methods on file 24/7.
  4. ACH is an option, but not a replacement as many customers won't put checking accounts on auto draft.
  5. Compte-rendu: When automated methods fail, how easy is it to identify upcoming expirations or expired cards? Can you confirm notifications were delivered? Are expiry reports available the way you need them- on demand or automatically distributed to A/R team, or both?
  6. Recurring billing technology - What happens when a transaction fails? It depends on the recurring solution. Here's some potential ways technology can reduce declines without customer contact:
  • Automatically attempt to reprocess on subsequent days, per card acceptance rules (which limit when and how many attempts)
  • Cascade to backup payment method on file, including ACH
  • If above fail, push to next month and charge current and past due on consecutive days (there are many ways to charge "last month"- it's important to identify the cycle desired for failures, plus how and when you want it to charge the current and last month, and then choose technology that matches). This is a critical area. Good questions are- will it charge past due and current due in a single transaction or multiple transactions? Is the recurring date fixed (always the first of the month for example) or flexible (anniversary billing 7 days a week)?
Leggi:  Où puis-je acheter (Antshares) NEO Gas?

Disclosure: I'm a reseller for related services.

Jaana Kulmala

Jaana Kulmala, founder of FirstOfficer.io SaaS analytics

Répondu il y a 204w

The card failures of all kinds are actually a big source of churn, especially when not taken care of. I often see 25% of churn being caused by credit card failures. It's not like the card would need to be lost or cancelled to fail, the reasons vary from no credit to just 'because'.

I didn't see any tools mentioned yet. If you are on Stripe, there are tools like http://churnbuster.io and https://bestunning.net that can take care of the dunning for you.

Matthew Caston

Matthew Caston, Partenaire, PCC. LLC | Angel | La graine | Conseil

Répondu il y a 237w · L'auteur dispose de réponses 644 et de vues de réponses 761.2k

You'll likely see a much higher percentage of expiration payment failures at volume. Likely 2-3x 1/36th amount. The good news is that this presents you with an awesome marketing opportunity - after all that's one of the reasons cards have expiration dates, so the banks and issuers can market to cardholders!

Expiring credit cards give you an opportunity to sell services related to a very real customer account message. Predictively, you can also use these messages (delivery fails, opens, clicks) to identify potential churn on an ongoing basis.

Anecdotally, and for what it's worth more than 70% of my cards expire in 2017 between March & Dec, even though there is a big gap in when the various accounts were opened.

Lascia un commento

Il tuo indirizzo email non sarà pubblicato. I campi obbligatori sono contrassegnati *

Questo sito usa Akismet per ridurre lo spam. Scopri come i tuoi dati vengono elaborati.