Bolt - New Wave Checkout

What is Bolt?

Bolt is checkout made easy (:

In reality, Bolt is actually a bundle of four different products that have way more value as a sum of parts and is pretty sick the more you think about it.

  1. Checkout UI: Bolt has built a streamlined UI for checkout.

    • Part of this is just performance optimizations: pre-rendering front-end components, quicker validation/load times, etc.

    • Another part is is easier usability through design (form design, autocompletion, one-click repeat checkouts, etc.)

    • The biggest advantages comes from fraud service/payment processing being coupled with Checkout UI data. Confidence in checkout behavior analytics allows for the elimination of elements that a payment processor would normally need as basic anti-fraud (entering credit card address and CVV).

  2. Anti-Fraud Detection: Bolt's primary feature is their anti-fraud protection. They absorb all first-order costs for chargeback fraud their platform misses (including friendly fraud). While covering chargeback costs is a nice sentiment, it’s more of a play to get merchants to trust their aggressive approach to fraud.

    They also claim that their anti-fraud is simply significantly better as a result of

    • coupling payments + checkout behavior (more data depth)

    • having a human approval system for flagged purchases on top of the platform (both from evaluating every blocked transaction from Bolt's end and a company override window)

    • allowing for minimized customization by having an iframe/plugin vs. API based approached (read: normalized data = better data = more efficient ML)).

  3. Payment Processor: As a payment processor, Bolt gets a couple of significant advantages:

    • Removes False Positive Optimization: One reason that merchants are so aggressive in flagging potential fraud are the second-order effects of potential chargebacks; mainly the risk of getting kicked off your existing processor or getting classified as a high-risk merchant

      • the costs associated with having a high-risk merchant account from a processor due to excess chargebacks (conservatively 50+ BPS on every transaction)

      • switching to a high-risk merchant processor also means a high rolling reserve requirement, which has a huge opportunity cost in a cash-flow poor business like retail

      With Bolt as your processor, the risk of being terminated for true fraud chargebacks is 0, allowing for a more aggressive approach. That being said, account termination from the merchant's back is always a possibility which Bolt still has to be mindful of.

    • Ease of Handling Disputes: Chargeback disputes have a lot of administrative overhead, both on compiling evidence for potential disputes as well as maintaining lines of communication with the processor on non-disputes (sorting chargeback codes, etc.). By sitting on top of the transaction, Bolt can manage this while only contacting the merchant if necessary shipping information is needed.

    • Outcome Visibility: Thinking about a traditional fraud detector built on top of checkout behavior, how would they know if they were wrong about a chargeback and adjust their model (either FP or FN)? It would involve:

      • someone from the company overriding the order in real-time (FP)

      • manually relaying the information (and the context behind the transaction) after being informed about a chargeback via processor or arrival bank (FN)

      • optimizing so that FNs never happened, creating a bunch of FPs (?)

      As a processor, Bolt can parse chargeback information as it arrives from the arrival bank before the merchant even gets it, significantly accelerating the feedback loop.

  4. E-Commerce Data: Bolt isn't doing anything proprietary here (handling guest user auth, caching for single sign-on, managing shipping address, discounts, order submissions, analytics, installation payments) but they still had to build an internal version of this that integrates with the rest of their checkout and with a company's backend. It appears their actually pretty heavily integrated with BigCommerce's SDK to handle this.

Rev. Model:

The unit economics for Bolt aren't standardized and depend on a couple of variables:

  1. Existing Payment Processor: Bolt's pricing is the typical 2.9x + .3 fee but they promise to match your existing processing fee. Since the market interchange fee is constant, Bolt makes much more money if you're switching from a "ease-of-use, high-service provider" (aka low fixed costs, high variable for your site) such as Stripe or PayPal as opposed to a shitty-service, cheap provider like Flagship (high fixed integration costs, low variable).

  2. Avg. Product Line Value: Bolt's per unit/fee is relatively simple; higher product cost = higher fee, which probably evens out for companies doing similar amounts of revenue at different per unit/prices. A good way of thinking about this as like an insurance premium; Bolt's risk for insuring chargebacks on Harleys is much higher than it is for golf clubs.

  3. How Bad Their Existing Solutions Are: Interestingly, Bolt also takes a small percentage of the revenue uptick a company has that's attributable to Bolt's checkout solution, at least as of 2018. It's not clear if this still exists as of 2020 and how this would be attributed to Bolt.

Why Now and Why Bolt?

Once upon a time, a small scrappy startup named Google tried to build a very similar product (Google Checkout, 2006) that fell flat on it's face.

Why Now:

  1. Higher Fraud Opportunity Cost: The cost of a lost customer, either because of being flagged as a false positive or a clunky checkout experience is exponentially increasing. There are a ton of macro trends around this; exponentially rising CPCs on paid channels because of static ad inventory, movement from direct to intermediaried traffic for traditional retailers, and increasing consumer choice.

  2. One Click Checkout: One of Bolt's UI selling points is one click checkout for existing retailers. Amazon held a (disputed) patent on this until 2017 so it literally wasn't possible for this exist

  3. Rise of Storefront Platforms: Bolt's "standardized" but extremely easy to integrate platform lends itself perfectly to storefronts like Magento, Shopify, WooCommerce, etc.. These platforms have had exponential growth the last 5 years

  4. Increasing Complexity: As the opportunity cost of lost customers increases, there has been a rise of number of 3rd parties that specialize in one part of the payments value chain: installment payments (Affirm), processing (Stripe), checkout analytics (Google/KISSmetrics), lifecycle marketing etc. Each additional system leads to more integration pains and also creates decision fatigue for buyers.

Why Bolt:

  1. Market Soft Spot: Bolt appeals to a very specific demographic; e-commerce businesses that have enough revenue for enterprise level spend (7-9 figure revenue) but haven't migrated off third-party storefront software. Most out-of the box fraud retailers (Signifyd, Sift, etc.) haven't optimized their products for this use case (plugins) and often involve really long integration cycles/costs or have limited full-scale product features.

  2. Instantaneous Bottom Line Uplifts: While Bolt (presumably) occurs a pretty heavy loss from human review of initially evaluating every potential case, the merchant will almost immediately see an uptick in revenue because false positives are pretty much wiped out. This leads to a more tangible outcome bottom-line upfront, instead of other 3rd party solutions, who might reduce the fraud rate initially, but have some of the bottom line effect offset by an uptick in false positives.

  3. Everything-in-One/Details: As cliche as it sounds, as a team dedicated solely for checkout, Bolt has nailed all the edge cases: design, built in installment payments, analytics dashboards (which includes conversion optimization), account recovery emails, etc. While other companies might be able to match or better parts of this functionality, it’s so much easier for the merchant to just click an on/off switch and not have to worry about anything except customer service after the user adds something to their cart. It’s the same argument for other ultra-bundled tools like Rippling.

    “There’s only two ways I know of to make money: bundling and unbundling” - Jim Clark

  4. Payments Review: Bolt’s centralized fraud review operation + chargeback guarantee takes away the need to hire people in house to review fraud. Hiring costs are really expensive if a company isn’t using contractors (training, benefits, etc.)


Competitive Landscape/Growth Opps:

Opps:

  1. Network: Bolt is targeting the long-tail of retail e-commerce; as companies such as Faire have shown (at the wholesale level), there is a ton of value in being able to create an Amazon-like recommendation network by aggregating a lot of the data behind low-purchase frequency, niche products. In fact, Shopify launched a standalone store discovery app the day I wrote this:

    • Standardized Login: One potential opportunity is to implement Bolt's single-sign on functionality across the entire network; sign up once, then be able to use one-click anytime you visit a company with a Bolt checkout page. It's actually a pretty blatant ripoff of what Fast is doing and a pretty hard technical challenge, but Bolt could also offer the additional promise of anti-fraud protection.

    • Recommendations/Discovery: Another option is to almost have an Outlook/Taboola-like recommendation engine across stores; display items from other Bolt network stores across the site or at checkout, which would drive low-cost incremental traffic to other retailers. It's an interesting future play if Bolt can build a dense enough base of customers.

  2. Dynamic Pricing: The combination of payments insights and really granular checkout behavior seems like they could be repurposed as indicators of intent. Again, with a pretty normalized dataset across all the customers, Bolt could use that aggregated data to either offer promotions in real-time to customers they detected were about to abandon checkout as well as way offer better lifecycle marketing tools (sub-segmented on IP, mouse movements, etc.)

Risks:

  1. The depth vs. breadth inflection point: Bolt's checkout might have better data depth, but they don't have a comparably large historical dataset to competitors yet. It's a little bit of a chicken and the egg problem; at scale, Bolt will probably surpass legacy players, but it needs to get to that point, which will be hard because their programmatic solution might not be as strong at first.

  2. Moving upmarket/Scalability: It's probably worth asking: what is the TAM of the type of retailer Bolt sells to? Even with an increasing shift from retail to e-commerce, it's not clear how many mid-size growing retailers on third-party storefronts there are. Part of the reason for e-commerce growth is the increasing ease of starting a storefront/social brand, which will mean vicious competition for acquiring traffic and a potential cap on revenue growth for most new retailers (short of the 8-9 figure range?).

    Bolt's focused on granular, normalized data also makes it hard to move upmarket towards bigger, more customized customers (there would be few reasons to switch to Bolt instead of Stripe).

  3. Competition: Even if Bolt is the best product on the market, the anti-fraud market is an extremely crowded space and has pretty high switching costs if you're using a third party provider (implementation time and loss of proprietary, historical data). E-commerce also doesn't seem to be like a super-high WoM space, so Bolt's GTM will have to involve:

    • preemptively identifying companies on growth curves that will land them in Bolt's target revenue range soon (using external data: fb ads, credit card, web traffic, etc.)

    • investing a lot on customer education and informing retailers how other fraud providers are taking a super conservative approach

    • competing aggressively on high-cost growth (i.e. heavy sales teams, massive SEM investment, etc.)

There's a couple of big competitors here, both in terms of upstarts and legacy competitors.

  1. Stripe: Stripe has a suite of products (Radar for anti-fraud, Checkout for checkout, and their core payments product) that can be bundled together to replicate Bolt's functionality. Stripe's target market is relatively different from Bolt's (developers) and their fraud data is probably less normalized because of the number of proprietary use cases that they have (marketplaces, retailer, etc.). However, if Bolt decides to move up-market or take an API-based self-serve approach down-market, it's pretty conceivable they would have to compete.

  2. Sift/SignifyID/etc: While these competitors might have as good of a product, they have beefed-up enterprise sales teams (for SignifyID specifically, a monopoly on SEO/SEM as well) and as incumbents, don't have to worry about customer education about their approach to fraud.

  3. Fast: Fast is taking an interesting approach to login/checkout; essentially, it's a web-wide network effect. You sign up for an account and enter information with Fast once, which auths you to one-click checkout for any merchant that has Fast as a checkout option. Fast's initial customer base is similar to Bolt's target market.

    At the end of the day, Bolt's goal is to increase checkout conversion for the largest possible base of users; even if Fast has the traditional anti-fraud FP problem, the increase in conversion from how much easier their product is to use could outweigh the conversion gains from Bolt's more complex checkout + false positive elimination.

  4. Shopify: Shopify's checkout/payments started out rather rudimentarily, but has gradually been improving; the checkout UI has gotten better, they've built seamless payments, and have implemented some level of fraud-protect (they cover all chargebacks that they flag as potentially fraudulent). Since there product is less specialized and there is no human review, it's probably not as good as Bolt, but as the default option (with presumably lower fees since Shopify has a million other ways to monetize merchants), it's going to be hard to get merchants to switch.

    While Shopify is just one of many storefront players, it's gradually taking away market share from competitors (though the storefront business isn't 0-sum).