How Predictive Analytics Demystify the iPhone 15’s Impact on the Secondary Mobile Market

Apple’s latest iPhone debuted early September and, as usual, started launching shipping shortly after that announcement. This time around, the release consisted of a standard and Pro version, each of which are available in two sizes, and various memory configurations. Apart from the long-rumored move to USB-C—likely a must-have feature for some—the iPhone 15 makes no giant leaps in terms of performance or features. From a consumer perspective, these advancements are largely iterative, and this upgrade cycle is less by hype than by contract and retailer incentives.

Although the launch prices on these newest models are not significantly different than in past years, these are still premium devices ranging from $800 to $1600. Understandably, consumers look for any way they can to bring that price down, and many are perfectly content to hand in their current device in exchange for credit towards their new one.

Naturally, securing these perfectly functional, still-premium trade-ins has become a priority for carriers and retailers.

Enter the Secondary Mobile Market

The secondary market—the purchase of pre-owned devices either at a consumer level or a business level—continues to grow, even faster than the market for new devices. There are, in fact, multiple reasons and implications for this trend.

Catalysts like COVID drove many people to buy additional devices as they were confined to their homes—devices for children’s schooling, remote work, entertainment—and often, these buyers looked to the pre-owned market, and have stayed in the habit since the pandemic has subsided. 

Another major factor pushing the preowned market forward is trade-ins, part of the reasons why iPhone and Apple launch events are more important than ever. As any savvy consumer knows, there are seemingly endless promotions from all US carriers all focused on recollecting you users’ old phones for discounts on newer models or credits towards other items or services. This subsidy model means high quality devices entering the secondary market at a higher rate than ever before, and each of these will all need home.

Before discussing the iPhone 15’s release specifically and how it impacts the secondary market, we’ll discuss how iPhone launches have played out historically with regard to the secondary market, as well as some factors that drive the iPhone market as a whole.

What Were iPhone Launches Like in Past Years?

In a pre-Covid world, these releases and the depreciation of Apple devices was extremely predictable, always following a very steady, easily understood curve.

This chart shows resale price of multiple iPhone models as a percent of original retail price over a 60-month period.

Of course, the depreciation was quite straightforward, accelerating slightly in relation to new releases. Of course nothing is as simple as it used to be, and big market changes have created macro-level impact on used device pricing on the B2B market.

What Large-scale, Macro-level Factors Impact the B2B Market?

The most influential drivers of the B2B market for preowned iPhones are as follows:

• Growing Demand for Refurbished iPhones
  Domestic demand for pre-owned, certified pre-owned, certified renewed, refurbished devices is growing. Consumers now have multiple platforms to shop for these pre-owned devices and rising direct-to-consumer prices have pushed B2B pricing up, driving some interesting trends in our data and altering the shape of the traditional depreciation curves.
• Supply Chain Issues
  A tired subject following years of Covid shutdowns, supply chain disruptions do still occur due to disasters, geopolitical events, wars, and so on, and these factors are always at the back of our analysts minds as they look at mobile resale data.
  
• Market Share
  Market share tends to ebb and flow—we see new contenders cropping up periodically, not to mention the never-ending battle between Apple and Samsung for smartphone supremacy. We certainly watch at the unit level what devices are making up the market because that will change trends and impact pricing downstream.
• Inventory Levels
  The basic principles of economics state that any time a producer dumps a massive supply of a particular good into the market, the value of those goods will be impacted. And this is another thing our analysts are watching for. On multiple occasions in the past, industry leaders have hit the panic button, so to speak, and released a wave of devices into the market leading to price disruption. Such events very much represent the human element of the market that can have observable effects at the macro level.
• Consumer Spending & Inflation
  Whether they’re buying new or pre-owned, each consumer has their own budget, and balances their mobile device purchases against other financial responsibilities. While phones have proven to be a priority in many people’s lives, the state of the economy and inflation will certainly impact how many devices are sold—especially new devices, which in turn affects the pricing of trade-ins. 
• Competitive Discounts
  Apple is famous for maintaining MSRP, and as a premium brand with a very strong product, they very rarely offer competitive discounts and promotions. This phenomenon remains on B-Stocks radar, however, as some OEMs may be able to pull business away from others through these programs, or bring new device prices down far enough to entice consumers away from buying pre-owned. 

What Small-scale, Micro-Level Issues Affect Device Pricing? 

The most influential determinants in an individual device’s value are as follows:

• Age
  Unsurprisingly, the simplest determinant of a phone’s value is its age. Consumers simply prefer newer upgraded versions over old versions.
• Model
  Usually, if a device model is new, it has more value in the B2B Market, however that does not necessarily hold for iPhones. For example, Apple now offers Pro models of each iPhone which may be more valuable than the base models of successive generations’ base models.
• Condition
  In a vacuum, an iPhone in good condition is more valuable and one in poor condition. An iPhone in excellent condition will command a much higher price than a comparable device showing even light use.
• Carrier Locked States
  Unlocked phones tend to bring higher prices than locked phones, firstly because they will function on any domestic network, but also because they will function overseas— where many second-, third-, and fourth-hand units wind up. 
• Seasonality
  Holidays and seasonal discounts can lead to shifts in consumers’ usual spending patterns, and thus changes the amount of inventory in the market, in turn impacting pricing. 

B-Stock’s Predictive Toolkit 

B-Stock has taken all of these macro and micro factors into consideration in building out our advanced machine learning algorithms. Our models can help each of our sellers accurately predict—at a highly granular level—the B2B and B2C market prices of iPhones as far as 16-20 weeks in the future. For example, it’s entirely possible to know what price a red, B-Grade, carrier-locked, 256GB, iPhone 13 will fetch on our marketplace two months from today.

This insight into secondary mobile market pricing is the value that B-Stock brings to OEMs, retailers, trade-in services, insurers, refurbishers, and third party dealers. With years of Market data and resale expertise, we can empower your organization to make smart data-backed decisions and optimize your margins. 

Data, AI, Machine Learning, and Predictive Capability 

B-Stock has been operating at scale and collecting B2B pricing data for all auctions across all phones since 2012. In fact, since 2017, we have sold 34 million units in 270,000 auctions, and have used every scrap of information we can to build our proprietary B2B mobile data—one of the largest in the world.

It’s this very data that enables deep insight through enhanced statistical techniques including outlier detection, interpolation, and machine learning.

Outlier Detection 

Outlier detection is the process of identifying unusual or abnormal data in the historical data set. For example, quick, seemingly random spikes can and do occur in all markets, but they are not always reflective of item

The blue arrow in the image above indicates a spike in the secondary market price of a C-grade iPhone 13 Pro Max—an obvious outlier that was perhaps caused by an extremely high erroneous bid or momentary market abnormality. Although the anomaly was immediately corrected, it could still throw off certain metrics, and therefore needs to be mitigated within the data set—something which B-Stock handles with ease. 

Interpolation 

In a simple terms, interpolation is a statistical technique to fill in missing data using existing data. We use techniques like spline interpolation, exponential moving average, K-nearest neighbors, and more based on the phone grade carrier locked State and many other factors.

Data before interpolation

Data after interpolation

Machine Learning

We predict the B2B preowned mobile prices at a granular level (model, grade, carrier, locked state, et .) using machine learning techniques like Auto ARIMAA, KNN, XGBoost LightGBM, and DeepAR.

The chart above graph shows historical and forecaster prices for iPhone 13 Pro Max unlocked grade-C. The purple line that you see is the historical B2B per unit price and the orange line is a weekly per-unit price forecast for the next 16 weeks, with the green area representing maximum variance

B-Stock MobileMind

This is B-Stock’s mobile pricing tool through which sellers will have access to all of this historical and predicted pricing and sales data. Through an easy to use interface you can filter your results by 12 different iPhone models as well as grades, carriers, and lock states, to see the market price trends and predictions. 

With this knowledge, you can better plan when to sell your inventory and understand what margins you can expect, and what trade-in value to offer to an end user.

As far as accuracy is concerned, our dynamic modeling process ensures that every week, the top performing algorithm’s output is reflected in MobileMind’s output. As an example, for different variants and conditions of the iPhone 12 and 13—popular trade-in devices at time of writing—average prediction accuracy ranged from 96-98%.

What’s Coming Next for B-Stock’s Predictive Analytics?

To further our goal of enabling intelligent secondary market recommerce operations, B-Stock plans to expand its MobileMind tool to include more models and variables that can affect device pricing. This work has already begun, with sights set on full pricing predictability for popular Samsung models in addition to iPhone.

Further, B-stock plans to feed more and more information into the AI and Machine Learning models with the goal of near-100% pricing accuracy. 

Despite B-Stocks trust in its predictive capabilities, there’s still a human intelligence element to our services, and great care is taken to bounce all predictions up against our team of experts to determine if the outputs truly reflect the market with which our clients are actually interacting, and what phenomena the AI might not be telling us.

What’s Next for Your Business?

B-Stock is working on delivering data solutions to help all sellers to drive their retail , trade-in, and insurance decisions by building this intelligence directly into their mobile resale programs. But while information is valuable, no value is realized until a business achieves a sale—that’s why B-Stock’s goal is to be a truly end-to-end solution—one that serves as both an integral advisory  partner and a highly effective marketplace platform.

Curious about how B-Stock can help your business get the most out of its pre-owned mobile inventory? Contact us today.

You can watch the full webinar on this topic here.