Run AppLovin from Your Dealership CRM

In app performance buying at national scale, and an unvarnished account of what a dealership would actually get.

AppLovin grew up as the machine that decides which advertisement a phone shows you inside a free application, and which advertiser pays for the privilege. It runs a large in app exchange, a mediation layer that publishers use to fill their inventory, and a bidding engine that has become the company's real product. That engine is why marketers talk about it at all, and it has been extended past its original job of selling application installs into web and commerce campaigns.

The inventory itself is video and interstitial placement inside mobile applications, a large share of it games. The unit is usually a full screen video or a playable, watched during a pause in something else, on a phone held six inches from a face. It is a very large amount of attention available at prices well below social.

For a store the question is not whether the audience exists. Everyone in your trade area with a phone is somewhere in that inventory. The question is whether an engine tuned to optimise toward a purchase completed inside an application can learn anything useful from a dealership, whose conversion is a phone call, a form, and then a person walking in three weeks later.

The case is efficiency and reach. This is one of the strongest performance bidding engines anyone operates, and inventory that most automotive advertisers never touch is priced accordingly. A store buying video impressions here would pay far less per thousand than the same store buying video on social, and the audience is not being carved out of a pool every other dealership in town is already bidding into.

The commerce expansion makes the argument sharper than it used to be. Campaigns that drive to a website rather than to an application store are now a real product, which is the only version of this platform a dealership could sensibly use, since nobody is installing a rooftop's application.

The honest objection is that the engine is hungry, and a dealership cannot feed it. Machine bidding of this kind gets good by observing many conversions quickly and adjusting. A store that produces forty or eighty opt ins in a month, geographically confined to a twenty mile radius, gives the model almost nothing to learn from. You end up paying for an engine you cannot switch on, and the results look like ordinary cheap video rather than like performance.

The second objection is radius. This platform is built for advertisers who buy nationally or globally, where the auction has depth. Narrow the geography to one town and the available impressions thin out, the auction gets less competitive in ways that do not favour you, and the cost advantage that made the channel interesting starts to erode. A dealer group across many markets has a much better version of this conversation than a single point does, and a single point is usually better served by the direct response channels described in our overview of dealership advertising.

Not started. The test attached to that status is stated so a dealer can check it rather than trust it: no adapter file exists in the ad gateway for this platform. No authorization URL, no credential block, no wire-up checklist, and no partial implementation waiting behind a status badge.

That is a materially different position from a registered platform, where an adapter and an ordered checklist already exist and the only open question is when the work is scheduled. Here the adapter would be the first thing written, which makes the request tally the whole mechanism rather than a way of breaking ties.

Access class is PARTNER, and that gate is the vendor's rather than ours. Advertiser credentials come through an application the vendor reviews, and a company whose bidding engine is its core asset is selective about who gets programmatic access to it. That review sits in front of the engineering regardless of how many stores ask, and pretending otherwise would set an expectation we could not meet.

Described as a design for a connection nobody has built, so the shape is explicit rather than assumed.

The third row is the one that undercuts the platform's own strength, and the section further down says so without hedging. An engine that learns from conversions cannot learn from conversions it is never told about.

An unbuilt platform can be described this precisely because the structure it would join already exists and already carries a live network.

There is one library inside LeadLocate that defines what any advertising platform must be able to do: hand us an authorization URL, exchange the credential for a usable token, list the accounts that credential reaches, deploy a campaign, pause or resume it, return spend and results, and prove on demand that the connection still works. Seven operations, one adapter file per network. Adding a platform is that one file written against those seven, which is why a request here asks for a scoped piece of engineering rather than for a research project.

Credentials are encrypted with AES-256-GCM before they are stored. Every action taken against any network is written to an append only audit log carrying the company, the network, the user and the detail, timestamped in UTC on purpose, because the table is shared across dealerships and a company local timestamp cannot be reconciled afterwards. Campaign objects record which credentials created them and re-resolve those same credentials later, so disconnecting an account does not orphan the campaigns built beneath it.

None of that is rebuilt per platform. It is the reason a hundred entry catalogue is possible without a hundred architectures, and the reason the honest answer on eighty of those entries is a queue position rather than a shrug.

Two ways to advertise through LeadLocate, and the store chooses which.

The first is the webXintel Ads Network: included, built and operated by our team from the selections you make in Leads Manager, and billed as a single line instead of a pile of platform statements. No application to file and no credential for you to hold.

The second is your own account. Connect the advertiser account your group already holds and campaigns route through it, on your billing, in your name, with your rates and history intact. For this platform neither path exists today, because no adapter has been written and the vendor gate has not been approached.

One constraint applies across the entire gateway and is better said here than discovered later: deploying, activating, pausing and reading spend is done by our team, not from a control on your dashboard. That is deliberate while the connection layer is young, and it is on the list to open up rather than something already shipped.

Four steps, two of them ours. A request from a store or a group that would genuinely fund the channel. A queue position earned against every other request in the catalogue. An adapter file written against the seven operations, with a credential block and a verification path. And a vendor reviewed application for advertiser access, on the vendor's timeline.

There is a prerequisite that sits outside all four. This inventory is video and playable, so the creative has to exist and it has to be built for a phone held vertically with the sound off. A recut of a television spot performs badly here and a still image performs worse. Creative production is not something we supply, and a store without a video asset should probably not be at the front of this queue.

On measurement, impressions and completion rates arrive quickly because the volume is enormous. What arrives slowly, if at all, is a defensible link between that volume and a sold unit, because the conversion happens off the phone and weeks later. Read the fast numbers as a check that the campaign is delivering, not as evidence that it worked. Nobody can promise the second thing, and a vendor who does is telling you something about themselves rather than about the channel.

There is no connection, so the gaps get named rather than blurred.

We have never made a call to this platform. No adapter file, no credential, no verification, no deployment, and nothing partly written waiting for a finishing touch.

We do not post conversion events to any bidding engine. No server side conversions API, no offline conversion upload, no pixel installation or management, and no software development kit inside anybody's application. That is the sharpest gap on this particular page, because a bidding engine deprived of conversion signal is simply a media buyer, and the whole reason a store would come here is the engine.

We do not emit a vehicle catalog or an inventory feed to this or any advertising platform, so dynamic vehicle level units are not something we could deliver even if the connection existed.

We do not export a hashed customer list for audience building, lookalike seeding or suppression.

We do not build mobile applications, and we hold no relationship with any mobile measurement partner. If a proposal you are reading depends on attribution inside an application, none of that machinery is ours.

Ask by name and attach the rooftop or the group. Requests are recorded against the platform they name and the tally is what decides build order, because a dealer with a budget ready to move is a better signal than any internal roadmap discussion.

Two details make a request here far stronger. Say how many markets you operate in, because this platform rewards scale and a multi point group is a genuinely different case from one store on one road. And say whether a vertical video asset already exists, since a channel that only accepts video is worth nothing to a store that has none.

Until then the managed path runs today and depends on none of this. Platform pricing is on the pricing page and does not change with the networks you enable.

Does LeadLocate integrate with AppLovin today?

No. The status is not started, and the test for it is that no adapter file exists in the ad gateway for this platform. There is no credential, no checklist and no partial code. The request count is what decides whether that changes.

Can I use my own AppLovin ad account?

Not through the CRM. Nothing has been written on our side, and access is partner gated, so advertiser credentials come through an application the vendor reviews. An existing relationship would remove the vendor half of that answer and leave the engineering half untouched.

What happens if I request AppLovin?

It is recorded against this platform by name and counted, and the count feeds build order across the catalogue. A person replies with a range or a reason. Requests from groups operating in several markets weigh more here than single rooftop requests, because the channel rewards scale.

Will the bidding engine optimise toward my leads?

Not through us. Optimisation of that kind needs conversion events posted back to the platform, and we post none. Without that signal the campaign is bought on delivery and video completion rather than on outcomes, which is a weaker version of what the platform is famous for.

Is this only for app install advertising?

Not any more. Campaigns driving to a website are a real product now, which is the only sensible shape for a dealership, since nobody is going to install a rooftop's application. The inventory and the auction dynamics are still inherited from the application world.

Can I target only my trade area?

Geographic targeting exists, but narrowing to one town thins the auction and erodes the cost advantage that made the channel interesting. Expect a very different economic picture at twenty miles than the national numbers a proposal will quote you.

Requests are recorded against the platform they name and the tally sets build order. Send your group, your markets, and whether a vertical video asset already exists.