Run Blis from Your Dealership CRM

Location advertising built to outlive the identifiers everyone else still depends on, and the visibility that costs.

Blis is a location based advertising platform that builds its audiences from patterns of real world movement rather than from an identifier attached to a device. Where a conventional buy follows a phone, this one describes a pattern of places and times and then finds the traffic that fits it. The advertiser buys a modelled group rather than a list of devices.

The design exists for a specific reason. The identifiers the advertising industry spent a decade building on have been steadily withdrawn by the platforms that issued them, and the supply that remains is thinner and less reliable every year. A vendor that never depended on them is not exposed to that decay in the same way, which is a genuine architectural advantage rather than a marketing position.

For a dealership this sits in the same part of the media plan as any other location buy, and the arguments for the channel itself are made on our geofencing advertising page. What is different here is the trade being offered: durability and reach in exchange for knowing less about precisely who was reached.

The case is that the targeting keeps working. A store that built its display strategy around following devices has watched that approach quietly degrade, and rebuilding it every time a platform changes its rules is a poor use of a marketing budget. Movement patterns in a trade area do not stop existing because an operating system changed a setting. For reach across a market, and for reaching commuters, weekend traffic and the areas your showroom actually draws from, a modelled location buy is a durable instrument.

The honest objection is visibility. A modelled audience is a black box by construction. You cannot inspect who was in it, you cannot verify the model that built it, and you cannot easily test whether it described the people you meant. The vendor is asking to be trusted about the input as well as the output, and on a channel where the same vendor also reports the results, that is a lot of trust in one direction. Ask what evidence exists that the model matched your market, and treat a confident answer with no method behind it as no answer.

The second objection is precision. Modelling trades exactness for durability, and for a dealership that wants a boundary around one named competitor, this approach is the wrong shape. It is built for reach across an area rather than for a tight polygon, and a store expecting the second will be disappointed by the first.

The third is that modelled audiences make exclusion harder. Suppressing your own existing customers, or your own staff, is difficult when the audience is described rather than enumerated, and this platform is not the tool for a campaign that depends on precise exclusions. Compare it against ordinary programmatic display before deciding which one your plan actually needs.

Available on request. The published test for that status is that the work is not started and demand decides the order, and this page does not stretch it. No adapter file exists in the ad gateway for this platform. There is no credential block, no authorization URL and no wire up checklist behind the scenes waiting to be finished.

Access class is PARTNER, which is the vendor's gate rather than ours. Credentials require an application the vendor reviews before issuing them. This is a platform with a larger footprint outside the United States than inside it, and its automotive tooling in this market is thinner than that of the specialists built for dealership advertising, both of which are worth weighing before a store asks for it by name.

The sequence therefore has two halves and only one is ours. We would schedule the engineering, decided by the request tally on this page. Somebody at another company would approve a credential, on a calendar we do not see. A date published here would be an invention of that second half, and the point of a status with a stated test is that it remains checkable by the person reading it.

A design rather than a description, because none of it exists.

The last row has a particular consequence on this platform. Vendors in this category will often offer to seed a model from an advertiser's own customer file, so that the modelled audience resembles known buyers. We do not produce that file for anyone, so that route is closed regardless of what the platform supports.

A page about an unbuilt platform can still be concrete because networks are not added here as bespoke projects. On a platform selling modelled audiences the operation that would need the most care is deployment, because the audience description going out is the campaign, and a payload that quietly loses half of it produces a buy nobody asked for and reporting that looks fine.

One provider agnostic library inside the CRM defines what any advertising platform must be able to do: hand us an authorization URL, exchange the code for a token, list the accounts a credential can reach, 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, answered identically by all of them. Nothing in that contract moves a customer record outward, on any network.

Twenty networks are registered against it and exactly one is implemented. The nineteen that are not return a refusal naming themselves, so a stray call fails readably instead of resembling an outage at the vendor. Blis is not among the twenty, which is a smaller and more honest claim than registered, and it is the claim the directory makes about eighty of the hundred platforms it lists.

Underneath the adapters sits infrastructure that already exists and would not be rebuilt for a new network. Credentials are encrypted before storage with AES-256-GCM. Every action against a network is written to an append only audit log carrying the company, the network, the user and the detail, stamped in UTC deliberately, because those tables are shared across dealerships and a timezone relative timestamp cannot be reconciled afterwards. Campaign objects record which credentials created them and re resolve those credentials later, so a disconnected account does not orphan the work built under it.

Two routes, and the store picks. The first is the webXintel Ads Network: included, built and operated by our team from the selections you make in Leads Manager, arriving as one line rather than a stack of platform invoices. That path is live and nothing on this page gates it.

The second is your own account. Where a network is built, you connect the account your store already holds, campaigns route through it, and the billing, the naming and the spend history stay yours. Disconnect and delivery falls back to the managed path.

Neither route reaches Blis today. On this platform the more useful observation is about who the buyer usually is. Advertisers here are typically brands and agencies working across markets rather than single rooftops, so a store asking for it directly should expect a conversation about minimums before a conversation about software. An integration would put the spend and delivery beside your other channels for comparison. It would not change the commercial shape of the relationship, and it would not make a modelled audience any easier to inspect.

The limit that applies across the entire gateway, stated on every page here: pushing, pausing and reading numbers from a network is done by our team, not from your dashboard. That is a deliberate constraint while the connection layer is young, and it is on the list to open up.

Two tracks. The vendor track is an application and a review before a credential exists, and probably a commercial conversation alongside it. Our track is an adapter file against the seven operation contract, a credential block, and each operation wired and checked against a live account. Ours is ordinary scheduled work. Theirs is not on our calendar.

The design question worth settling early is how much of a modelled audience definition can be represented in a campaign payload at all. Where a definition is built inside the vendor's own console and cannot be expressed in an API call, an integration ends up deploying budgets against audiences a person assembled elsewhere, and an honest implementation says so on the screen rather than implying the CRM built them.

On measurement, expect a reach read rather than a response read. Results in this category are not reliably attributable inside twenty four hours. Decide before the campaign runs how many people in the area you intended to reach, how often, at what cost per thousand, and what you would expect to see in your own showroom traffic and direct enquiries over the following weeks. That is a statement about how to read the channel, not a forecast of visits, leads or sales, and nobody can produce that forecast honestly for a modelled buy.

The gaps below are the ones that matter on this platform specifically.

We do not supply a customer file to seed a modelled audience. Vendors in this category commonly offer to build a model from an advertiser's own buyers, and there is no code in our stack that exports such a file. That route is closed here whatever the platform supports.

We do not upload a hashed audience or export a customer list to this or any advertising platform, which also means we cannot suppress your existing customers from a campaign.

We do not measure store visits and no part of our stack observes a device or resolves a location. Any movement or visit figure attached to this channel is the vendor's own.

We do not run a server side conversions endpoint or an offline conversion upload. Neither exists anywhere in our stack, so closing the loop from an impression to a sold unit is not something this connection would do.

We do not install or manage a tracking pixel. Site side measurement is whatever you already had.

We do not retrieve leads from this platform. Lead form retrieval is implemented for exactly one network and this is not it, and a display impression collects no opt in.

We do not emit a vehicle catalog. No unit level feed goes to any advertising platform from us.

No campaign has ever been deployed to Blis from the CRM, because there is no code that could do it.

Ask, with your store attached, and say whether you want reach across a market or precision against a named place. That single distinction decides whether this platform is the right tool for you or the wrong one, and it is the question a vendor conversation is least likely to put to you first. A modelled audience is built for the first job and it is a poor fit for the second.

Requests are recorded against the platform they name, and the tally is what decides build order, because a request from a store ready to spend is a better input than an internal roadmap discussion. You will hear back from a person with a range or a reason.

Expect the reply to be honest about scale as well. A single rooftop asking for a platform built around multi market advertisers may be better served elsewhere, and being told so is worth more than being sold to. The managed path runs today regardless, and pricing for the platform is on the pricing page.

Does LeadLocate integrate with Blis today?

No. The status is available on request, whose published test is that the work is not started and demand decides the order. There is no adapter file, credential block or wire up checklist for this platform in the ad gateway.

Can I use my own Blis account?

Not through the CRM, because nothing exists to reach it. Advertisers here are more often agencies and multi market brands than single rooftops, so a store asking directly should expect a commercial conversation about minimums before anything technical.

What happens if I request Blis?

It is recorded against this platform specifically and counted, and the tally is the input to build order. A person replies with a range or a reason, and the reason will note that credentials need a vendor reviewed application, which is a step outside our control.

What does targeting without device identifiers actually mean for me?

It means the audience is described rather than enumerated. The upside is that the approach does not decay each time a platform withdraws an identifier. The cost is visibility: you cannot inspect who was in the audience or verify the model that built it.

Can you seed the model with my own customers?

No. We do not export a customer list or a hashed audience to any advertising platform, and no code in our stack produces such a file. If a vendor offers to model an audience from your buyers, that file would have to come from you, not from us.

Is this a good fit for fencing one competitor's lot?

Generally not. Modelled audiences trade precision for durability and are built for reach across an area rather than a tight boundary around a named address. If a single competitor is the target, this is the wrong instrument even where it is available.

That distinction decides whether this platform fits your plan. Requests are counted against the platform they name and the tally decides build order.