Don’t go the way of Blackberry or Nokia. Using datascience to understand how your product matures can save you from corporate death

It is not the strongest of the species that survives, nor the most intelligent that survives, rather that which is most adaptable to change.

Every new product that’s introduced to the market has a maturity curve. It starts with the initial euphoria of fast growth, the glamour of being the new belle of the ball and the adoration of users.

Everything eventually comes to an end

But pretty soon, the shine wears off. The product begins to lose steam. It has a steady, faithful user base, but eventually newer entrants to the market take over, and the process repeats.

Everyone knows this. But the biggest issue frustrating product managers is the lack of data to foresee. There isn’t a lot of product specific data out there to help product managers make data-driven choices. Most design gurus (like IDEO) often advise product managers to get out there and test it (which is management speak for getting your own damn data ☺). But even IF you manage to do that, the sample size is going to be ridiculously small.

So what’s a product manager gonna do?

We subscribe to Alon Halevy’s thesis that an abundance of data can be unreasonably effective at making predictions. We don’t think that you need to follow a theorized model of the product maturity curve. Instead, you can (and should) use the data at-hand as proxies for the insights you need.

Product maturity curves can be inferred from numerous market signals. Category benchmarking is an excellent way to predict or infer your own product curves using data from a basket of similar products.

Here’s a great way to do this: Lets take, for example, the Blackberry Z10 and the Apple iPhone 5S. Both products were released in 2013, look somewhat similar and share similar performance specs.

Here’s sample JSON output for the Blackberry from the Semantics3 API:

{
 “cat_id”: “9359",
 “ean”: “0802975662012",
 “model”: “Z10",
 “offers_total”: 18,
 “weight”: “136077.71",
 “category”: “Unlocked Phones”,
 “name”: “BLACKBERRY Z10 UNLOCKED”,
 “description”: “Product Description The phone’s original packaging… (visit site URLs for full description)”,
 “updated_at”: 1412359036,
 “price”: “249.88",
 “width”: “129.54",
 “upc”: “802975662012",
 “brand”: “BlackBerry”,
 “created_at”: 1376532374,
 “sem3_id”: “0kuAckKGZ60e48gWUWSw4I”,
 “color”: “BLACK”,
 “features”: {
 “Display Resolution”: “1280 x 768 pixels”,
 “Input Device(s)”: “Multi-touch”,
 “Mobile Broadband Generation”: “4G”,
 “Communication Features”: “Mobile Email client, Internet browser”,
 “Sensor Resolution”: “8 Megapixel”,
 “Capacity”: “1800 mAh”,
 “Product Type”: “BlackBerry smartphone”,
 “Technology”: “Lithium ion”,
 “Data Transmission Operating Frequency”: “LTE 700/850/1700/1900",
 “Max Operating Temperature”: “95 °F”,
 “Diagonal Size”: “4.2\””,
 “Cellular Messaging Services”: “MMS, SMS”,
 “Instant Messaging Services”: “BlackBerry Messenger”,
 “Navigation”: “A-GPS/GLONASS receiver — Simultaneous GPS (S-GPS)”,
 “blob”: “3G/4G LTE connectivity; BlackBerry 10 OS; Dual-core 1.5 GHz processor; 4.2-inch touchscreen display (1280 x 768); 16 GB storage + microSD expansion”,
 “Data Transmission”: “GPRS, EDGE, HSDPA, HSUPA, HSPA+, LTE”,
 “Body Color”: “Black”,
 “Focus Adjustment”: “Automatic”,
 “Wireless Interface”: “NFC, Bluetooth 4.0, IEEE 802.11a/b/g/n”,
 “Digital Zoom”: “5",
 “Personal Information Management”: “Calendar, synchronization with PC, calculator, reminder, stopwatch, alarm clock”,
 “Connector Type”: “Micro-USB Headset jack — mini-phone 3.5 mm HDMI”,
 “Navigation Software & Services”: “BlackBerry Maps, AT&T FamilyMap”,
 “Min Operating Temperature”: “32 °F”,
 “Bult-in Memory”: “16 GB”,
 “Sensors”: “Accelerometer, ambient light sensor, proximity sensor, magnetometer, gyro sensor”,
 “Included Accessories”: “Stereo headset, USB cable, power adapter”,
 “Form Factor”: “Touch”,
 “Supported Digital Audio Standards”: “WAV, WMA, AAC, PCM, Ogg Vorbis, AMR, MP3, FLAC, AC-3, AAC-LC, AAC +, OGG, eAAC+, MKA, MIDI, M4A, eAAC, WMA 10 Pro, WMA 9 Pro, WMA 9",
 “Processor Core Qty”: “Dual-core”,
 “Color Depth”: “16.7 million colors”,
 “Supported Flash Memory Cards”: “microSDHC — up to 32 GB”,
 “Band”: “WCDMA (UMTS) / GSM 850/900/1800/1900",
 “Application Software”: “File Manager, DataViz Documents To Go, Microsoft Direct Push, Microsoft Exchange ActiveSync, Foursquare, BlackBerry Protect, Weather, Adobe Reader, BlackBerry Browser, BlackBerry Calendar, BlackBerry Remember, Story Maker, BlackBerry Connect for Dropbox, BlackBerry Hub, BlackBerry World”,
 “Run Time Details”: “Talk ( WCDMA ) — up to 600 min Standby — up to 312 hrs”,
 “Messaging & Data Features”: “Microsoft Word support, Microsoft Excel support, Microsoft PowerPoint support”,
 “Operating System”: “BlackBerry 10 OS”,
 “Additional Features”: “TTY compatible, DLNA Certified, software updates FOTA (Firmware Over The Air), multitasking, BlackBerry Balance, BlackBerry Magnify”,
 “Camera Light Source”: “LED light”,
 “Phone Functions”: “Speakerphone, voice control, call timer, conference call, flight mode, voice dialing, vibrating alert”,
 “Lens Aperture”: “F/2.2",
 “Other Features”: “Picture stabilizer, video recording, video stabilizer, tap to focus, Time Shift mode”,
 “LTE Band”: “LTE Band 4, LTE Band 17, LTE Band 5, LTE Band 2",
 “SAR Value”: “1.5 W/kg”,
 “Compliant Standards”: “HAC(Hearing Aid Compatible)”,
 “Mobile Services”: “Video Call, YouTube, AT&T Hot Spots, myAT&T, BlackBerry Push Service”,
 “Supported Digital Video Standards”: “MKV, VC-1, AVI, MOV, M-JPEG, XviD, MPEG-4, WMV, ASF, 3GP, H.264, H.263, M4V”,
 “Graphics Accelerator”: “Qualcomm ADRENO 225",
 “Clock Speed”: “1.5 GHz”,
 “User Memory”: “16 GB”,
 “Video Recorder Resolutions”: “1920 x 1080 (1080p)”,
 “RAM”: “2 GB”,
 “Supported Social Networks and Blogs”: “Twitter, Facebook, LinkedIn”,
 “Service Provider”: “AT&T”,
 “Integrated Components”: “Digital camera, 2nd camera, digital player, GPS receiver, Wi-Fi hotspot, GLONASS receiver, voice recorder”
 },
 “height”: “10.16",
 “length”: “66.04",

Similarly, here’s a sample JSON output for the Apple iPhone 5S from our API:

{
 “cat_id”: “9359",
 “ean”: “0885909727490",
 “model”: “ME310LL/A”,
 “operatingsystem”: “iOS 7, upgradable to iOS 7.1.2, planned upgrade to iOS 8",
 “offers_total”: 118,
 “category”: “Unlocked Phones”,
 “name”: “Apple Iphone 5s 32gb”,
 “description”: “LATEST MODEL, MOST FEATURES OFFERED THUS FAR FROM … (visit site URLs for full description)”,
 “updated_at”: 1414453781,
 “price”: “734.99",
 “upc”: “885909727490",
 “brand”: “Apple”,
 “created_at”: 1380111937,
 “sem3_id”: “1ob9w2uqXYSyaM8CgaAk2i”,
 “color”: “GOLD”,
 “features”: {
 “Shipping Information”: “View shipping rates and policies”,
 “blob”: “APPLE IPHONE 5S 32GB; AT&T; ALL ORIGINAL ITEMS SEALED; PLUS EXTRAS”
 },

Through our database, we track a number of market signals for e-commerce products, which are used to calculate our proprietary ProductRank, which in turn seeks to measure the absolute popularity of a product in our vast catalog against everything else in the database.

High ProductRank (i.e. a rank of 1) means that a product has very high volumes of sales, while a low ProductRank (i.e. ∞) means that a product has low sales overall. Taken over time, the decay of a ProductRank can be used as a proxy for product maturity curves.

The new product maturity curve, derived NOT from theory and/or modelling, but from pure data. Source: Semantics3 API

As this chart makes abundantly clear, although it’s pretty evident that the iPhone 5S dominates the Blackberry, what’s also interesting is how its ProductRank decays over time. The iPhone goes through several phases when its ProductRank jumps up, then decays with multiple “sparks”. But more importantly, it takes much longer for its ProductRank growth to flatline.

The Blackberry, on the other hand, has far fewer “sparks”. Not only that — apart from the couple of times its rank spikes, its ProductRank decays pretty rapidly. It’s pretty evident that the Blackberry Z10 wasn’t a very popular product, with both the CMO and COO being ousted over its failure.

By using both products as extremes, you could actually get a very reasonable estimation of how well your product might perform over a similar product cycle, if it belongs to the same category, and shares similar product characteristics. If you get data over an even longer time line, and do it for a higher number of products (i.e. even more data), your estimations would only get more accurate.

Curious to know more? Our database has all of this information, captured over time. You can access all of this via our API and create your own product maturity curves using our data at Semantics3