Kelly and I were sipping coffee at Digital Dealer, greeting participants, and speculating on how the ultimate online buying experience would come to pass. Presenters had talked about Amazon, obviously, and the recent opening of a Hyundai digital showroom on Amazon Autos.
A while back, I organized the various offerings into categories like: online platforms where multiple dealers may list their inventory (basically lead providers) versus eCommerce plug-ins to be placed on individual dealer web sites.
One key variable was whether the site actually holds inventory, i.e., is a dealer, not just a technology play. Carvana, for example, or Shift. Increasingly, what I notice is that the good technology either evolved from a dealership, or – I found this intriguing – they will buy a dealership to serve as a test bed.
Your rapper name is a top twenty dealer group plus a digital retail system.
Roadster came from a concierge buying service which, as far as I know, they still operate. A2Z Sync came out of Denver-based Schomp group. The Gogocar people operate a Kia dealership. This brings me to the next level of dealer technology tie-ups, those where big dealer groups choose an online retail solution and commit to it.
We philosophically do not believe that software development is our expertise. Instead, we’d prefer to partner with third parties – Craig Monaghan
That prediction is … continuing the consolidation megatrend, we will see dominant groups taking the lead in online retail, but unable to master the technology on their own. This is what I call the “Kodak syndrome.” Incumbent leaders are not agile enough to ride a paradigm shift. This means not only the dealer groups, but the traditional software vendors.
I expect to see the Sonics and Asburys of the world buying up the digital retail people, absorbing their talent, and denying access to their competitors. I characterized this as a “land rush” in the earlier piece. Direct to consumer is the final frontier.
In today’s post, I add to the copious literature on technical debt with a discussion geared to my audience of F&I entrepreneurs, and extend the metaphor into organizational design. What I noticed, writing Maturity Model, is that software development is rich in models and metaphors that apply outside the trade.
Technical debt is incurred when software developers take shortcuts, usually because they are under time pressure. This debt is accrued in program code, but it must eventually be paid off with real money, just like the debt on your balance sheet. Here is a brief discussion of how that works.
Someone once insisted that my team “just code IF State = TX, and get on with it!” Not on my watch. We will categorize Texas, and then we will add other states to the category as we discover them. For example, the category might be “waiver GAP states,” or “spousal consent states.”
Operating or back-office issues, often related to IT, are recurring concerns for strategic buyers. Problems with IT underinvestment have proved to be ordeals during many integration efforts.
If you go down the road of IF State = TX, then in short order you will have code with IF <list of ten states> do this logic, and for <five of them> also do this other logic, but for <three of the ten> do this instead and for <the other two> do both of those things.
Congratulations, you have saved forty hours of programmer time versus stubborn Mark Virag and some academic exercise involving categories. Now you are married to this gnarly decision tree, and you will be debugging it forever. The technical term is Big Ball of Mud.
One warning sign of technical debt is the “cut and paste” approach. If your developers implemented the latest dealer, provider, lender, product, or state by copying code from the one before, then I guarantee you have technical debt.
Any developer worth his salt will, instead, make the copied code into a reusable method. Developers are trained to do things the right way and, in my experience, only take such shortcuts because the boss told them to.
Why not cut and paste, if it gets the job done? Because, if there are any bugs in the copied code, now they’re multiplied and scattered throughout the code base. You will have to spend programmer time to fix each one separately, as they are encountered over time.
I could go on with examples all day. The point I want to make is that technical debt is real money. You may go “quick and dirty” this week, and save $5,000 of developer time, but you will be paying those same developers later when they have to fix the bugs.
You may reasonably decide that you are a little short this month, and take a loan from the invisible bank of technical debt, but you should do so consciously. Don’t fool yourself that technical debt is free. I have provided an example here in the form of TILA box, something my F&I readers will understand.
Now that I have that off my chest, let’s discuss investment decisions. For example, if you’re a startup and strapped for cash, you may choose to pile up technical debt because it’s off balance sheet and may be the only kind of financing you can get.
Of course, no one actually thinks about it this way. What they tell their developers is, “just keep patching it until we’re profitable and you can overhaul it later.” You may even sell the company, rickety software and all, if the acquirer fails to do proper diligence.
When I was doing international software search for BMW, our due diligence guy in Munich was Dr. Dettweiler. We would find some software that looked pretty on the outside, and then the doctor would fly in and discover it was all a façade, like a movie set held up with sticks.
McKinsey specifically warns against acquiring a company with a big ball of mud in the back office. Like process maturity, this is a concept that goes beyond software development.
In my time as a consultant, I have designed an organization or two, and it’s a lot like programming. You have to have the right boxes on the org chart, with the right procedures and job descriptions, kind of like designing objects that will respond to business events (except they’re people).
Organizational debt is caused by the same kinds of things that cause technical debt. For example:
The structure worked fine ten years ago when we had one-tenth the number of people.
It was never actually “designed” to begin with, but we reorganize ad hoc every other year.
The structure is based on specific people instead of job functions.
There are processes for which no one is actually responsible, so things “slip through the cracks.”
Fortunately, people are remarkably resourceful. They will create their own procedures and informal networks. Good people can prop up a bad organization, like those sticks holding up the movie façade, but they can only hold out so long. Sooner or later, they will start to slip – and customers will start to notice.
Now I feel like I really am writing a pitch for consulting services. Call now! Free reorg with every digital transformation. Seriously, though, my point is that organizations can harbor technical debt just as software can. This is why I am a fan of formal methods like ISO certification and, yes, professional organizational design.
More broadly, I am starting to notice that software development concepts – like process maturity, technical debt, iteration, and agile teams – are applicable throughout the enterprise. We’ll explore this further in an upcoming post.
Last week, I presented a single-column format for menu selling on an iPhone, with the glib recommendation to let analytics determine the sort order. Today, I will expand on that. Our task is to sort the list of products in descending order of their relevance to the current deal, which includes vehicle data, consumer preferences, and financing terms.
This sorting task is the same whether we are flipping through web pages or scrolling down the mobile display. The framework I present here is generalized and abstract, making the task better suited to automation, but ignoring the specific F&I knowledge we all take for granted. I’ll come back to that later.
For now, let’s assume we have six products to present, called “Product One,” and so on, and four questions that will drive the sorting. Assume these are the usual questions, like, “how long do you plan on keeping the car?”
That answer will be in months or years, and the next one might be in miles, but we are going to place them all on a common scale from zero to one (I warned you this would be abstract). Think of using a slider control for each input, where the labels can be anything but the range is always 0.0 to 1.0.
Next, assign four weights to each product, representing how relevant each question is for that product. The weights do not have to be zero to one, but I recommend keeping them all around the same starting magnitude, say 1 to 5. Weights can also be negative.
For example, if there’s a question about loan-to-value, that’s important for GAP. High LTV will correlate positively with GAP sales. If you word that question the other way, the correlation will still be strong, but negative. So, now you have a decision matrix that looks something like this:
Yes, we are doing weighted factor analysis. Let’s say that, for a given deal, the answers to our four questions are, in order:
[0.3, 0.7, 0.1, 1.0]
To rank the products for this deal we simply multiply the decision matrix by the deal vector. I have previously confessed my weak vector math skills, but I am certain that Python has an elegant way to do this:
Product Two ranks first, because of its affinity for high-scoring Question Four. Product Four takes second place, thanks to the customer’s response to Question Two – whatever that may be. By now, you may have noticed that this is the setup for machine learning.
If you are blessed with “big data,” you can use it to train this system. In a machine learning context, you may have hundreds of data points. In addition to deal data and interview questions, you can use clickstream data, DMS data, contact history, driving patterns (?) and social media.
If not, you will have to use your F&I savvy to set the weights, and then adjust them every thousand deals by manually running the numbers.
For example, we ask “how long will you keep the car?” because we know when the OEM warranty expires. Given make, model, and ten thousand training deals, an AI will dope out this relationship on its own. We will do it by setting one year past the warranty as 0.1, two as 0.2, etc. We can also set a variable indicating how complete the manufacturer’s coverage is.
Same story with GAP. Give the machine a loan amount and a selling price, and it will “discover” the correlation with GAP sales. If setting the weights manually, set one for LTV and then calculate the ratio for each deal.
Lease-end protection, obviously, we only want to present on a lease deal. But we don’t want it to crowd out, say, wearables. So, weight it appropriately on the other factors, but give it big negative weights for cash and finance deals.
I hope this gives some clarity to the analytics approach. In a consumer context, there is no F&I manager to carefully craft a presentation, so some kind of automation is required.
Followers of my Twitter feed know that I have lately been looking at mobile apps, to see if anyone can present protection products on an iPhone. I wrote about this three years ago and, according to my informal survey, the field is still open.
I don’t think anybody has a good way to present a menu on a consumer web site, much less an iPhone.
Not only is the iPhone a restrictive form factor but we must assume that the customer, not an F&I person, is operating it. We would like to apply our Best Practices for Menu Selling, but the app must be able to apply them on its own.
For example, if we want to retain the package concept with the carefully chosen payment intervals, we can use an accordion control. I proposed this for a client once, in an F&I context, but it doesn’t make sense for consumer use.
No, the best way to “present all the products, all the time,” is simply to make one long column with everything in it. The iPhone presents challenges, but there are offsetting advantages. We can show fifteen products in one column, and the customer has his leisure to scroll through them.
I prefer scrolling to swiping for a few reasons. In the prototype shown here, we have the obligatory vehicle photo. After the first scroll, that’s gone and the screen space is devoted to products.
The prototype shows monthly prices for the vehicle and the products. This assumes the finance process is settled, and the app can choose products matching the finance term. Touching any of the products will open up a full page with details, coverage choices, and a “sales tool” as in the earlier article.
I recommend using analytics to determine the sequence of products in the column, and even to A/B test the format of the product blurbs. I have in mind a few different formats:
Text with graphic and price, as shown here.
No price ‘til you open it.
Lead with the sales tool.
I discuss analytics here, but I am not a fan of the full “ownership survey.” Of the eight standard questions, maybe you can sneak in one or two elsewhere in the process. Apart from that, we’re counting on data points found in the deal itself.
I also think “less is more” when confronting the customer with choices. As you can see in the mockup above, there must be no complicated grades of coverage (or deductible). If you’re configuring the app for a specific dealer, you may want to filter some options out of the dealer’s product table.
Depending on who’s managing the app, the products themselves may be rethought. If you want to offer chemical, dent, key, and windshield as a combo product, then that’s a single choice. Alternatively – since we have unlimited column space – you can offer each one individually. What you do not want is a product having fifteen different combinations.
Coming back to my informal survey of mobile apps, and the workflow given here, I believe there are already good examples of vehicle selection, credit application, trade valuation, and payment calculator. Menu selling has been the only missing link, until now.
Feeling quantitative again today, so … suppose you have an F&I Director, or a menu trainer, or somebody, and their goal is to move product index from 1.0 to 1.2 over some time period. To keep the numbers simple, let’s say the variable comp component is $10K.
One way to do this is to say that 1.2 pays $10K and the current performance, 1.0, pays zero. This makes sense, right? Why pay for no improvement? This only works, however, if you place a cap on it. As the salesman, I could come back and say, fine, if the two points of product index are worth $10K to you, what happens if I hit 1.4? Are you willing to pay me $20K for that?
Most people resist the idea of capped pay plans. Mathematically, you are making a linear relationship between compensation and performance, and you should be willing to honor that relationship up and down the line. The problem here is that the line is too steep.
So, let’s try a shallower slope. Once again, 1.2 pays $10K, but this time the zero point is 0.0. That means the current performance, 1.0, still pays $8,333 and the salesperson doesn’t go hungry unless the index actually falls all the way to 0.0. Obviously, this plan is too weak.
The weak plan may be desirable if your sales force is really counting on some of that money, and the “variable” is not as variable as advertised. It also protects the company on the up side. In this example, I can achieve a 1.5 index and it only earns me an extra $2,500 (on top of the $10K).
Now we have examples of two pay plans, one too strong and one too weak, as in the story of The Three Bears. The key to making the pay plan just right is to observe that the zero point is arbitrary. In the papa bear case, too strong, we set the zero point at 1.0, the current performance. In the mama bear case, we set it at an index of 0.0.
Recall that a line on a graph is determined by two points. One point is fixed by the target and the bonus (1.2, $10K) the other point is set by where you place the zero (z, $0) and between them they determine the slope:
If you’re making this chart right now remember to place the independent variable, product index, on the x-axis. All that remains is to compute the y-intercept of your line. The point where the bonus is zero, z, is the x-intercept. A little algebra gives the y-intercept as:
Now we are ready to start plotting. This time, let’s split the difference and set the zero point at 0.5. This seems to be just about right. Comp for standing still, 1.0, is $7,143; for 1.5, only $14,286, and our guy doesn’t starve until 0.5.
If you have this set up in a spreadsheet, you can tweak the zero point until you have the desired amount of exposure for both parties. Below is the chart for my “goldilocks” line, with the mama bear case for comparison.
We always give careful thought to the target, but sometimes neglect the slope of the payoff line. Next week, we will talk about two-dimensional pay plans, combining product index and PVR.
I concluded Car Dealer Megatrends with the clear and present dominance of consolidated groups, which I like to call the Best Buy phase. Today, I will indulge in a little futurism, and explore the Amazon phase. In the Amazon phase, it will be possible to buy a new car enitrely online and have it delivered.
By 2025, experts estimate 30-40% of car sales will be online. The high end of that range is from Mark O’Neil. Used cars are easier to sell online, witness Carvana, Vroom, and Shift, but new cars will be there too. An estimated 25%, and that’s only seven years away.
The industry is rapidly solving problems like pricing and trade valuation. The only challenge people still talk about is the test drive. Carvana solves this with its seven day return policy, and Shift will bring the car to you for a test drive.
“The current dealer model is not a dying breed,” Benstock said. “It’s dead. It’s absolutely dead.”
I will order a new BMW sight unseen, because I know the product and I trust the manufacturer. Their online configurator is awesome, and I really would press the “build and ship as shown” button, although the process isn’t quite there yet. We’ll come back to BMW later, but for now let’s assume a test drive is required.
The tension between Best Buy and Amazon centers on a practice known as “showrooming.” This is where you sample the product at Best Buy, interrogate the Best Buy sales associate, and then turn around and order the product from Amazon. Amazon even makes a clever app you can use to scan product codes while you’re in Best Buy.
As auto retail moves into its Amazon phase, I can imagine the same challenge for dealers. You have invested in a monument to your manufacturer’s brand image, where customers can sample the product and then go order it online.
I had been pondering the showrooming challenge for a while when I ran across this piece in the Wall Street Journal. Nordstrom is opening stores with no stock, where shoppers can try on clothes and accessories, and then have them delivered.
It will contain eight dressing rooms, where shoppers can try on clothes and accessories, though the store won’t stock them.
The Nordstrom story reminded me of the old “catalog showrooms” operated by mail order retailers like E.L. Rice and Service Merchandise. Ironically, this was the last gasp of mail order, put out of business by brick and mortar retailers – including, ultimately, Best Buy.
All of this goes to show that, in the Amazon phase, showrooming and fulfillment can be disconnected. Where the customer goes, to test drive and learn about the vehicle, does not have to be the dealership or even affiliated with the dealership. This opens up a world of new possibilities.
I can think of several applications for standalone test drive centers. For instance, suppose a manufacturer wanted to enforce its ideas about how to present its vehicles, and also – since this is the Amazon phase – protect its own position online.
Were it not for U.S. franchise laws, manufacturers would run their own retail outlets. In Europe, they have company stores, where ideas about brand image, sales training, and product positioning do not depend on a network of autonomous dealers.
An OEM test drive center would bypass the dealer network (or complement it, if you prefer). It would be staffed by salaried, factory-trained product experts with no other objective than to educate customers in the finer points of their company’s vehicles.
There would be minimal inventory, attractive video displays, simulators, and samples of paint and fabric. No transactions would take place, but there would be plenty of Wi-Fi bandwidth and gourmet coffee for the online shoppers.
As I said, this is just one scenario. The new techniques of digital retail will create untold opportunity for dealers willing to adapt. Our exploration of the Amazon phase has just begun.
I felt like doing something quantitative, so this week we look at seasonal adjustment factors. Everybody always talks about SAAR, and you probably know that it stands for “seasonally adjusted annual rate,” but what does this really mean?
Well, suppose it’s March 2017, and you are wondering what total sales will be for the year. The industry sold 1.55 million vehicles that month so, if you multiply by twelve months, you might estimate 18.6 million for the year.
You would be wrong, though, because March is always a strong month. Here are the estimates produced by the simple “times twelve” method, relative to the actual total for 2017, which was 17.2 million.
Using data from Fred for the five years 2013 through 2017, and converting everything to a percentage, you can see how March always overestimates the year’s results. Each year’s dots are a different color, though it doesn’t really matter which is which.
Some months are highly variable, like September. Not a good gauge of anything. Remember to distrust any SAAR figures published in September. April, oddly, is a tight group and bang on the annual rate. April 2018 sales were 1.4 million, so a good guess for the year is 16.8 million.
Taking an average across the five years, we find that March, May, August, and December each overshoot the annual rate by roughly 10%. Finally, we convert these percentages into monthly adjustment factors.
Instead of multiplying last month’s sales by 12, multiply by the monthly factor to predict the year’s total. Of course, we have more data than just a single month. We can also look at cumulative sales since January. For example, do a quarter of the year’s sales occur in the first quarter?
No. It takes a while to make up for the weak January and February, and then the actual historical cumulative pace slowly comes into alignment with the idealized linear cumulative pace. I made that chart, too, but it’s not pretty. That’s enough quantitative stuff for this week.
I was asked recently to opine on this topic, which I do today with some reservation, for I can see the venerable four-column menu approaching its sell-by date. The image shown here is a MenuVantage prototype from 2003. Don’t get me wrong. As I wrote here, this is still the best tool for the traditional setting in the F&I office … for as long as that setting prevails.
Best practices for menu selling split into two broad categories: those that are good for selling, and those that are good for compliance. I will present them in that order.
Every product appropriate to the transaction type and “car status” of the current deal (i.e. Used Lease) should appear in column one. Some menu systems use deal templates, making it easy to select the proper layout every time.
The home court advantage in the F&I suite is that you can do a four-column menu, and there is a professional there to present it.
For most systems, column one automatically drives the layout of the accept/decline “waiver” form. This is best practice for compliance, and it’s good selling too. Why have a product that you only present on special occasions?
The practical limit for products in column one is six, maybe eight, so choose wisely when laying out the menu template. Using bundles will allow you to squeeze in more products. I generally don’t like bundled products, as I wrote here, but this is a reason to use them.
Every menu should include a second, longer term, with the correct APR for that term. There is a charming story about this in Six Month Term Bump, plus a downloadable spreadsheet. Twelve months is overkill, and likely to raise an objection.
The amount of product you can finance without changing the monthly payment is given by this formula. Without doing the annuity math, a good approximation is: base payment times five.The monthly payment in column four should be roughly $30 more than the base payment without products. That way, you draw the customer’s attention into the menu without a big price barrier. Likewise, payments should increase in small increments from right to left across the bottom of the menu.
Obviously, the increments will be larger for more expensive deals, say 10% of the base payment. This is easy to do, if you are manually setting up each menu. It takes a little more planning to do this with templates. You can either tweak the individual products at deal time, or you can set up a different template for highline vehicles.
For example, offer the platinum VSC coverage in column one and the gold in column two. By the way, do not reuse the VSC coverage choices (like gold, silver, and platinum) as your column headings. That’s an obvious source of confusion. Finally, your menu system should feature sales tools and custom content for each product, like the famous depreciation chart for GAP.
I have a few more recommendations, related to compliance. If you already have a good grasp of unfair and deceptive practices, you can skip this part. Be warned, though, that consumer watchdogs and regulatory agencies are looking over your shoulder.
The chart below (and the pull quote) is from the National Consumer Law Center. You can tell that the dealer in green is using a menu system with a fixed markup over dealer cost. The dealer in red is certainly making more PVR but he is also courting a federal discrimination charge.
Menu trainers like to say, “present all the products to all the customers, all the time.” They might add, “at the same price.” The NCLC report goes on to show that minority car buyers are systemically charged more for the same products. Some dealers simply don’t allow the F&I manager to vary from the calculated retail price. In states like Florida, that’s the law.
Giving F&I managers the discretion to charge different consumers different prices for the same product … is a recipe for abuse.
The menu should display the price of each product, not just the package price. Some turn this into a selling feature by also showing the price as a daily amount. It makes a good layout to have the most expensive product at the top, with prices descending down the column.
All of these measures require some kind of audit trail. I have seen some very strong systems that track exactly what was presented, by whom, when, for how much, and whether the price was changed. At a minimum, you should collect the customer’s signature on the waiver form, with all the products, their prices, and your standard disclosure text.
Next week, I will resume writing about the brave new world of flow selling, self-closing, and predictive analytics. We may find that many of these practices – especially regarding compliance – are still relevant.
Around the turn of the century, I was helping RouteOne to build their now-ubiquitous credit system. Then, I moved on to aggregation models for the “I” side of F&I. It was a lot of work.
We had to develop scores of unique interfaces for lenders and product providers. We had to develop deal calculation engines, and then reverse engineer each DMS so our payments would match. There were no automated sources for finance or product rates. We had to walk ten miles in the snow …
Today’s eCommerce startups have it easy. All of the key tasks are supported by readily available services, leaving the entrepreneur to focus on user experience and dealer support.
When I started writing about this space, the key challenges were price negotiation and trade valuation (and the test drive, but I’ll cover that in a later piece). Today, you have reliable online trade valuation from Kelley, Trade Pending, and others. Price negotiation can be handled through chat or one-price, generally on used vehicles.
You can have payment calculations, including incentives, from MarketScan, provider networks from PEN or F&I Express, and finance networks from RouteOne or Dealertrack. Everything in this paragraph is an API, not to mention passing data from your eCommerce platform into the corresponding dealer system. Finally, even the old faithful DMS now exposes a variety of databases, like inventory.
A few months ago, I described the role of venture capital in driving process change. I think this eCommerce ecosystem is equally important. Entrepreneurs can enter the space at a very low cost, relative to ten years ago, and meet most of their requirements through interfaces.
We flew east out of Panama City, and I looked down on the faceted green hills of the Cordillera de San Blas, perhaps for the last time. In the sky were statistically similar puffs of white cumulus cloud. Naturally, I was thinking of fractals.
I had spent the past few days coding technical analysis indicators in Python, which reminded me of coding same in C#. This, in turn, reminded me that although the TA community talks a lot about geometric repetition, we have yet to invent a single fractal indicator, much less a trading strategy.
I write my trading strategies in C# on the MultiCharts platform. Its procedures for time series data look a lot like the vector-oriented syntax of Python. Here is how to do Bollinger bands in each:
df[price].rolling(length).std() * devs
I have to admit not having much intuition about vector operations. Matrices and summations just look like for loops to me – clearly an obstacle to the proper appreciation of Python. I have worked with SAS and SYSTAT, though, so Python at the command prompt seems natural.
What I noticed with the Python exercise is that the classic TA indicators were designed with an iterative mindset, reflecting the programming languages of the day – Sapir’s theory, again – and so I imagine that the reason we have no fractal indicators is that our language can’t express them.
Here are some basic things I would expect from a fractal-oriented programming language: