Clairvoy: A New Product Idea

[joe.pistell]

Dear Clairvoy,
Here is an idea that could become a significant addition to your platform. Hope this helps….


Product Concept Summary:
I believe you may have the data and resources to forecast a car dealers total sales AND forecast what are his TOP 10 vehicles most likely to sell this week. Machine Learning can look for patterns of activity that commonly precede a sale.


Product Benefits.
This product is a NEW product. Car dealers don’t care who they sell the car to as long as it’s sold. If this concept is true, then hundreds of new opportunities will spring from this.

The output of this work is spoken in a language ALL dealers understand and will use daily. For example:

• A GSM that has a ‘grinder’ of a customer who wants another $500 off. The GSM looks to his Clairvoy HOT SHEET and sees that this unit has 25 unique viewers looking at it this week, it’s ranked #11 in most VDP views and Clairvoy has graded this unit at 90% likely to sell this week. The GSM holds his ground.
• The buyer of new or used cars looks to the Clairvoy HOT SHEET to bring in similar units.

Also, by tracking all of the online activity, it may be possible to forecast the dealer’s sales. If the forecast is accurate enough to be actionable then all dealers will want forecast updates because they all run on quotas!

Bonus! This concept can happen without a DMS connection.


Like what you're doing!
Best Wishes,
-Uncle Joe

 

[Ryan Everson]

I know CarStory by Vast recently released a tool that sounds similar to what you're describing

https://www.carstory.ai/

 

[joe.pistell]

thnx Ryan, good find.
Does CarStory include the dealer's site traffic data? I don't see it. From my research as a marketing director, dealer site traffic data should contain the highest quality signals, which may produce the more accurate forecasts.

 

[Ryan Everson]

thnx Ryan, good find.
Does CarStory include the dealer's site traffic data? I don't see it. From my research as a marketing director, dealer site traffic data should contain the highest quality signals, which may produce the more accurate forecasts.

Yeah I'm not sure but since their other product, CarStory Market Reports, involves placing a script on SRP and VDP pages I would think it shouldn't be too hard for them to utilize those signals if they aren't already doing so.

 

[joe.pistell]

anyone here from @CarStory able to speak to this?

 

[chad.bockius]

Joe,

Ryan's correct. We launched a sales prediction product late last year called CarStory Insights. We use machine learning and AI to predict when every car on a dealer's lot will sell.

Your point about online engagement is spot on and Ryan's correct. We use the CarStory Market Report (which is free for dealers) to capture demand signals. While we also use demand data coming from marketplaces, there is no question that dealer site engagement is the strongest and purest signal.

In addition to online engagement, we are also analyzing supply and demand for similar vehicles, feature preferences in a local market and market position. Beyond the vehicle, we also take into account a dealer's past performance. We believe that predictions need to be dealer specific vs. just looking at a piece of inventory in a local market.

In about a week we'll be rolling out a new capability called the Price Explorer. This feature gives you the ability to run what-if pricing analysis on each vehicle. Not only will this help with day-to-day pricing management, but it also helps with situations like the one you described above. As always you'll be able to see your competition and when their vehicles are predicted to sell at current prices.

Here is a quick video demo that shows the Price Explorer and Competitive detail:

Also, the product is free to try for 30 days. If you're interested, you can sign-up here: https://www.carstory.ai/insights/

Hope this info helps.

-Chad

 

[john.quinn]

I dunno.... I gotta say... (In no way a direct comment on either of the products or services mentioned above -- which I have not researched)

The enthusiasm around "new" analytics reminds me of the 7-year cicada cycle -- every few years, we all get excited about our newfound potential to effectively predict the future. We putz around for a while, realize there's really not anything very actionable, maybe keep it around for a while, and place it at the top of the list to jettison as soon as the boss says "tighten the belt."

Not that trying to predict consumer behavior isn't an age-old effort -- like alchemy (with similar results). Perhaps someone here will break the cycle? That'd be cool

 

[joe.pistell]

I believe if Analytics is accurate, then, they may have the capability to PREDICT something (e.g. total car sales, most likely to sell, etc).

If analytic's can't predict, then I'd use it as a guide or reference.

 

[Alexander Lau]

I know this much, there must have been something in the IP of predictive analytics company https://www.automotivemastermind.com that motivated IHS to buy them for around $400M = http://www.autonews.com/article/201...ires-automotivemastermind-in-bid-for-behavior. That's not chump change man.

I knew a few of those German guys, from their days at Mercedes-Benz Manhattan. I once interviewed with them.

 

[chad.bockius]

At the most basic level, machine learning and AI-based algorithms replicate what we do every day. Our minds take in a variety of information, and we try to make the best decision possible. Unfortunately, we can only synthesize so much information. We also aren't very good at looking back at past decisions and figuring out where we went wrong. AI changes this.

Everything in this space is accelerating:
- We have access to more data than ever before
- Computers can process all this information faster
- The algorithms are more powerful

The exciting thing is that these new solutions can learn. The most famous example of this is AlphaGo. In 40 days this system taught itself to play the ancient Chinese game of Go and become the best in the world. If you recall, it took IBM's Deep Blue ten years to beat a chess grandmaster.

The concept of ongoing improvement and learning is one of the most exciting advances with AI. These systems won't be 100% right all the time. No system is. Not even a human. But they have a distinct advantage over us. They learn and get better, much faster than we ever will.

I don't think we are at a place where we need to hand the keys over to the machines. But I do believe that arming yourself with this new source of information will result in better, more profitable decisions.