Presenting Products on Shopping Sites

I asked what seemed to be an unexpected question at the F&I conference. Every provider I have spoken with is keen to meet consumers earlier in the shopping process, but few seem to have engaged their technology partners.

All of you are working with technology partners, like menu systems and provider networks, and I would like to know what discussions you have had with them about presenting your products to online consumers.

This is what I had in mind when I wrote that system vendors should be aligning their B2C plans with product providers. I imagine that some network or menu system, somewhere, is soliciting input from its provider council, but I have not found any evidence of it.

Protection_DriverI did see one demonstration of a consumer experience designed to dovetail with an in-store menu presentation. It was impressive, but still the design was driven by the system vendor. I have to believe that providers have some untapped insights. I know that was the case with finance sources when I formed a committee to tackle online credit, back in the era before Dealer Track.

Online Car Buying, Continued

In the previous instalment, we covered about half the required tasks in our buying process. Not by coincidence, we began with a precedence diagram and we stopped at the topic of navigation. I began my inquiry with precedence because I wanted to see if there was perhaps a better sequence for the tasks.

Most of my readers are familiar with the traditional dealer-based process, which is basically sequential. Our web-based, customer-driven process does not have this constraint. The process is constrained only where tasks have prerequisites.

Online PERTWe could enforce a sequence, but I am not a fan of artificial constraints. A passive approach is better, wherein the customer will choose her own path through the process. This means the site must manage the state of the work in progress. In this article, I will refer to the work in progress as a “deal.”

Before we proceed, let’s digress for a moment and assume that my passive approach is incorrect. You will have coded, all for nothing, a robust deal class and a set of tools for managing its state. The customer, it turns out, was happy to be led through the process. She has no need for a feature like, “keep my lease terms the same but swap in an upgraded vehicle.”

On the other hand, if you take a rigid approach and you later need to add such a feature, then you will have real problems. You are better off building a proper framework up front. Adding constraints to a flexible process is a lot easier than going the other way.

I envisage a process where the customer can use our tools for shopping, and then register a free account when she wishes to save the deal. Graphic elements on the site will show how complete the deal is, and the customer can decide when she’s ready to send it to a dealer. She can save multiple deals, which may share attributes.

Trade In and Trade Up

I have talked to people who consider trade valuation to be just as problematic as pricing the sale vehicle. This is not my area, but I can see the need for an accurate valuation. Unfortunately, there is no good way to appraise a car online. A visit to the dealer may be required.

Like the earlier discussion about inventory, much depends on the business model. If the site is operating on behalf of an individual dealer, then there’s a need for precision on every vehicle. If the site will handle a high volume of trade vehicles, but its primary purpose is to sell new vehicles, then an accurate average will suffice.

What I would suggest is to present the Kelley Blue Book value, and be prepared to honor it unless the customer is materially wrong in his condition assessment. The difference between “good” and “very good” for a five-year old Ford Focus is $250. Also, if the dealer is using a provisioning system like vAuto, then there’s an opportunity to develop an interface and let it make the decision.

Again, this is not my area. I am standing on the general principle that, if there is a dealer system, then we can interface with it. The business model dictates whether there is one relevant system, or many. We’ll see this again when we come to the credit application.

Choose Your Payment

I named this section “choose your payment” because that’s how our customer is likely to use the deal structure tool. Most shopping sites support only a forward calculation, proceeding from the sale price to the monthly payment. Lending sites provide a simplified “rollback” calculation, which allows the customer to start with the desired payment and work backward to the sale price.

Unfortunately, these sites omit local taxes, fees, and incentives. These items can make a big difference to the deal structure, and including them in a rollback is computationally complex. I have some experience with these calculations, from my time at MenuVantage. If you have to choose, skip the rollbacks and, instead, allow the customer to clone and save multiple deals.


An individual dealer will keep track of the relevant incentive programs. If more than one is available, he will choose the best combination. To do this online, you must link to a service like Market Scan. Taxes and fees mean another data feed, or two, not to mention rates and residuals for one or more lenders. You can see how this is easier if we restrict our scope to one OEM and its captive.

All of these features are available in the dealer’s Desking system. Some sites, like AutoNation, have the advantage of standard systems across the dealer network. Other sites have Desking capabilities built in. I am trying to keep this article general, and not make assumptions about the dealers. Here are some suggestions:

  • Enforce a standard Desking system at the dealer level, and interface with it.
  • Engage a Desking system vendor to work with the site, including local tax and OEM data.
  • Support multiple Desking systems and try to develop a generic interface.
  • Build your own Desking system, including data feeds (not recommended).

Fees, like an Etch fee, might vary among dealers. Consolidator sites have an advantage in harmonizing fees.

Standard functions for any Desking system include calculating a loan, calculating a lease, comparing two deals, converting a loan to a lease, and experimenting with different rates and terms. Much Desking functionality, like the venerable “four square,” is sales oriented and not relevant to our online consumer. I would really like to see an objective way to compare two deals, like calculating an IRR, or presenting the TILA values.

Obtain Financing

Here again, I am going to delegate the work to a dealer-based credit system, like Route One. These systems are the most advanced among dealer systems, in terms of being plug-in ready for consumer sites. See my earlier article, Claiming Space on Consumer Sites, for more explanation.

A site controlled by an OEM will have an easier time with rates, incentives, and credit application

Since we are doing indirect lending, we can’t postpone the “choose a dealer” task any longer. Our customer must apply for credit via a specific dealer. The fastest way to do this is with an interface that routes through the dealer’s credit system, selects two lenders using rules from the dealer’s profile, and sends them the application.

Obviously, the F&I Manager would prefer to route the credit application, and we can offer “assisted routing” as an option, but we don’t want to impede the customer. The F&I Manager is not going to complain about approved apps appearing on his dashboard.

At this point, our customer is ready to print and sign the contracts at home. Signing the contracts electronically is not a requirement, although we will utilize some e-contracting infrastructure. I had a hand in developing this infrastructure for Provider Exchange Network, and I am certain it will stand up in a consumer context.

Remember that our design objective is to empower the customer and save time in the dealership. Physically signing the papers is not time consuming. There are other reasons to do e-contracting, which we will leave to the dealer systems.


I hope this article has given you some useful ideas about online car buying, as distinct from shopping. I set out to show that the entire process, as we know it in the dealership, can be transferred to the consumer space.

The basic technique I propose is to take the functions of systems already in the dealership, interface them, wrap them, and organize them into a consumer web site. I also presented a rough outline of how navigation and deal management would look in this context.

Car buying sites, as they develop, will have different business models. They can be categorized according to their relationships with the participating dealers. Because the site interacts with the dealer systems, the nature of this relationship has a profound impact on site design.

For example, a site controlled by an OEM will have an easier time with rates, incentives, and credit application. A site controlled by a dealer group will have an easier time standardizing systems, fees, and trade valuation.

Our design goal was to improve efficiency by making the customer autonomous, but this is not a dealer averse process. It is a flexible process that allows the customer to engage with the dealer as much or as little as desired. Modern dealers should have no difficulty with flexible processes and empowered customers.

Design Concepts for Online Car Buying

In my last article, I made a distinction between online car shopping and buying. Shopping is an iterative process in which the customer settles on a vehicle, with only a rough idea of the price and terms. For many customers, this is sufficient. They will then proceed to the traditional dealer-driven process.

In this article, I will present tools and techniques to support online car buying.  As a software designer I feel that, fundamentally, this is something we should be able to do. If the tools exist in the dealership, then there’s no reason why they can’t be on a consumer site. This was the subtext of my earlier article, Dealer Systems in the Consumer Space.

I agree that customers should visit a showroom for help in choosing a vehicle, and I agree that two-thirds of them (that’s the survey figure) will prefer to do the paperwork there. On the other hand, why shouldn’t the customer have the option to complete some or all of the paperwork at home? I envisage a process driven by the customer, with support from the dealership.

We’ll start with a simplified process model based on task precedence, and then we’ll discuss the design requirements. I have a more detailed model based on data flows, but the precedence model is easier to work with. The flowchart below shows the required sequence of tasks, given their prerequisites.

Online PERTThe first thing you notice about this diagram is the number of deadlocks. You can’t be certain of the finance rate until you get an approval, and you need it to structure the deal. So, the sales desk must use a floor rate or do a pre-approval. Even in the dealership, this is an inherently iterative process.

Real-life business processes are never as clean as you would like. Still, I propose to streamline this one and hand it over to the consumer. Since we’re working from a precedence diagram, I’ll try to remember and discuss data requirements as we go along.

Choose Your Car

The customer may search the inventory of a given dealer, all models of a given OEM, or all dealers participating in an online search platform like Auto Trader. For our purposes, AutoNation is a giant dealer site, not a platform, because they control the inventory. This is an important distinction.

An OEM site may list all the vehicles they make, but must still search dealer inventories to find a specific unit. GM’s Shop-Click-Drive starts with selecting a dealer (an obvious concession to dealer pressure) whereas does a proximity search just as any platform would.

This question of whose inventory to search, and when, is not so much a process decision as it is a business model decision. That’s why I left “choose a dealer” off the diagram. For example, the difference between Vroom and any other used car platform is that Vroom takes the unit into inventory. Prices and policy are set by a single entity, instead of an aggregation of competing dealers.

The business context may vary, but the software requirements do not. Inventory search means model and trim selection, photos, zip code, etc. I seem to be permanently cookied in Farmington Hills, no matter whose search platform I use.


In the earlier article, I discussed adding an inventory search plugin to an online lending site. All of the high level tasks in our diagram already exist as components of various web sites and dealer systems. You could conceivably deliver the entire process using service orchestration.

Find Your Price

Once the customer lands on a vehicle, she has also landed on a price. A special task to disclose the price is needed because other dealers will undercut our online price. This is the most challenging part of our process, and I have written previously how I feel about it.   Here are some solutions:

  • Use no-haggle pricing. This is more practical with used cars, but new car dealers will come around eventually.
  • Join TrueCar’s conspiracy to “destroy the industry.” What the last guy paid for his MKZ is not the dealer’s intellectual property.
  • Do the haggling online, or by phone. For that matter, the customer can visit the dealership, take a test drive, haggle, and then go home to finish the paperwork.

Remember, our objective is to make the process available to the customer online. Nothing stops her from contacting the dealership for help.

Millennials are starting to express the fact that they’d like to do more shopping online. They’d like to connect that shopping to a buying experience…that needs to be quicker and more transparent.

Depending on the site’s business model, there are a number of ways to handle pricing. The trick is finding one that works for the customer, without alienating the dealers. By the way, an OEM is in a much stronger position to reform pricing than TrueCar is. The quote above is from Bill Fay at Toyota.

Protect Your Investment

Protection products are strangely absent from the shopping sites. When you put the process into precedence terms, selling a Vehicle Service Contract jumps right to the top. As soon as the vehicle is specified, even before it’s priced, you have enough information to quote a VSC.

Whatever sales approach you wish to implement – like, citing the total cost of ownership – is available early in the process. This applies to Road Hazard, Maintenance, and Appearance Protection, to name a few. My favorite sales tool for GAP must wait until the deal structure is known, but nothing stops you presenting the product at this stage.

Whose products to present is another question. If this is an OEM site, our choice of products is fixed. Let’s design for the worst case and assume a platform with a variety of dealers. In this case, we must store a profile for each dealer, as we would for a menu system, indicating the dealer’s slate of products.

Which sales approach is best for the online consumer? I’d be lying if I said I knew, because no one knows. I have some experience with menu presentation, and I am pretty sure that’s not the answer. I have also done some work with expert systems.

For the purpose of this article and our Millennial online customer, I am going to recommend a passive approach. In my next article, I’ll expand on passivity as a design theme. What it means for this task is simply exposing links to the products, including information about the chosen vehicle, and allowing the customer to educate himself. Also, bear in mind that as a commerce site, you should be doing A/B testing.

This question is really about navigation, and how to manage the customer’s experience. That’s where we’ll resume, next week.


Car Buying

Claiming Space on Consumer Sites

In my last article, I described a technique for extending traditional dealer functions into the consumer space. I used the metaphor of a land rush, with the extension feature claiming territory on one or more consumer web sites. This is part of a generic strategy, to update any dealer system and make it more relevant to the evolving world of online auto retail.

In this post, we will survey the consumer space, looking for opportunities to insert the various dealer system extensions. My plan is not to predict the future of online retail, but to show where these “plug-ins” may be placed today. Prediction only comes into play because we want to know how durable the relationship will be.

Sunnyvale Toyota is one of many individual dealer sites featuring the eLEND credit plug in. If you’re not clear on the plugin concept, take a look. Technically, this is an Iframe that passes control to the extranet formerly known as Dealer Centric. The site itself is by, and the dealer has chosen to delegate the credit process to eLEND.

B2C Diagram

I really like this technique, but that’s not the question. Credit systems are well represented on dealer sites, using a variety of techniques. The question is – where are the protection products? Pricing and presenting products on a dealer site is technically easier than doing a credit app. This is an obvious place to extend menu systems and provider networks.

A dealer site is not a good place to start desking, because the site is too far down funnel, and the process depends too much on pricing. A key advantage for the dealer’s desking system is the ability to include local taxes, fees, and incentives. We’ll talk later about how and why desking should be part of the customer’s online shopping experience.

A dealer site is also not a good place for opt-in CRM or, at least – as I wrote last week – not as good as an OEM site. People move around too much. Another thing to consider with dealer sites is that they will increasingly be dominated by shopping sites and consolidators.

You would think that specialist car shopping sites would offer the most functionality to their customers – and you would be wrong. That’s because they don’t own the inventory, and they don’t control the process at the dealership. It’s hard to initiate a credit app or sell a service contract on behalf of a heterogeneous dealer population. Auto Trader offers a link to a direct lender.

Even consolidator sites, like Sonic and AutoNation, leave the heavy lifting to their stores. They are better able to standardize process and systems than the other shopping sites, but they still have to work with different franchises and different captives. There is opportunity here for dealer systems to differentiate themselves by offering consumer-site features.

You would think that specialist car shopping sites would offer the most functionality to their customers – and you would be wrong.

Used car consolidators, like CarMax and Vroom, have much more flexibility. It is theoretically possible for these sites to offer a complete buying experience online. To me, that includes credit, desking, and products. Vroom wants to be “the Zappos of online car buying.” This brings me to that futurism topic I promised to avoid, and our next category … car buying sites.

I like to make a distinction between shopping sites and buying sites. Shopping is a nonlinear process, with rough ideas and estimates. Just look at all those asterisks on the shopping sites. Buying requires things to be done in a certain sequence, with exact values.

A shopping site is all about selecting and pricing a vehicle from inventory, and then turning over the lead. This means a variety of dealers with a variety of systems, and little opportunity to extend functionality toward the customer.

They may say “buy a car online,” but very few actually attempt a complete buying process. Vroom and ShopClickDrive are exceptions. Both use Route One for credit processing – and nothing for protection products. Buying sites, as they evolve, will be the Promised Land for dealer system extensions.

I was going to make a separate category for research sites, like Edmunds and TrueCar. For our purposes, these are the same as shopping sites, above.

Calculators and credit apps abound on the auto finance sites, and some of them are quite good. This is where you can “roll to amount financed,” for example, with current loan rates. I see untapped potential here, for desking, CRM, and product links. Plus, here’s an out-of-the-box idea – an inventory search plugin for the banks. It would look something like this, below. Bank of America already has the window. It’s just missing the car!


I include OEM sites even though they’re poor candidates for dealer system extension. The challenge with OEM sites is that they already have proprietary relationships with F&I partners. The Buick site takes you right to Santander, and the products on the Hyundai site are from Safe-Guard. Also, an OEM must consider all their dealers, so they can’t favor one system over another.

Your best bet here is to develop a generic interface, and then compete on the depth and breadth of your OEM integration. Dealer Track does this in the service department. If you know of anybody doing it in the front office, please let me know.

The table below summarizes the opportunity for extending dealer systems into the consumer space. There is no “official” taxonomy of auto retail sites. I chose these categories because they work well with the generic strategy.

Sites Table

Whether you are planning a strategic move into the consumer space, or just a new feature, I hope you find my analysis useful. You can’t beat the land rush if you don’t have a map.