FourKites Alternative: When SMB Freight Brokers Should Self-Host Visibility

Every quarter or two, a broker sends us the same email. They had a great call with FourKites, the demo was sharp, the network coverage map looked impressive, and then the quote came back at $32,000 a year on a two-year commit with $18,000 of implementation on top. Their question is always the same: is there a real FourKites alternative, or is this just the cost of being a freight broker in 2026?

The short answer is that there is a real alternative for a meaningful slice of the market, and it is not another visibility vendor. It is self-hosted freight tracking that runs on the ELD APIs you are already paying for. Most SMB brokers we talk to already write a Geotab, Samsara, or Motive check every month. The data that powers half of what FourKites sells you back is already flowing into your carriers' fleet systems.

This post is the honest version of that conversation. What FourKites actually costs at SMB scale. What the self-host path looks like when you build on Geotab or Samsara APIs your carriers already have. Where FourKites still wins. And the volume threshold where the math flips one way or the other.

What FourKites actually costs at SMB scale

FourKites does not publish pricing, and the salespeople will not give you a number until call three or four. We have seen enough quotes to give a reliable range. For a broker doing 30 to 100 loads a day, with a single TMS integration and standard carrier coverage, a FourKites contract in 2026 lands between $25,000 and $55,000 per year on the platform fee alone.

That is the floor. The $25,000 minimum is real, and we have not seen a quote below it in two years. FourKites is not interested in the under-25-load-a-day broker. Their sales motion is built around mid-market shippers and 3PLs, and the smaller broker accounts are tolerated, not pursued.

Implementation is a separate line item. Budget $12,000 to $35,000 depending on how many systems they have to integrate with. A McLeod TMS implementation is on the lower end of that range. A custom-built TMS or a Tailwind shop with EDI handoffs lands in the middle. Anything with multiple carrier portals or non-standard data flows pushes higher.

Contracts are typically two or three years. One-year deals exist but the per-year cost goes up by 20 to 35 percent if you refuse the multi-year commit. So the realistic three-year total cost of ownership for an SMB broker on FourKites runs between $90,000 and $200,000, depending on volume tier, implementation scope, and how hard you negotiated.

For a broker doing 40 loads a day at $180 average gross margin, that is $2.6 million in annual gross profit. A $30,000 FourKites bill is 1.15 percent of GP. Not catastrophic. Not free either. It shows up on every quarterly P&L review for the next three years, and somebody in operations gets asked whether it was worth it.

The freight visibility cost question is really a sizing question

FourKites is not overpriced. It is priced for a different customer than the average SMB freight broker. The platform was built for shippers with $50 million plus in annual freight spend, and 3PLs running multi-modal global networks. At that scale, the per-load cost of FourKites is in the $0.20 to $0.80 range, which is a rounding error against the total freight spend.

At SMB broker volumes, the per-load cost of the same platform is closer to $2 to $5. That is not a problem with the product. It is a problem with the fit. If you are paying $4 per load for visibility on a load with $150 gross margin, you are giving up 2.7 percent of margin per load just for tracking. Some shippers absolutely require it. Most do not.

This is the same shape as the Project44 conversation we covered earlier this month. The enterprise visibility platforms are built for enterprise freight. There is a sub-segment of the broker market where they make sense, and there is a much larger segment where they do not, and the dividing line is not subtle once you do the math.

The data you already have access to

Here is the part most brokers do not realize until somebody walks them through it. If your carriers run Geotab, Samsara, or Motive, the GPS data, the engine status, the HOS clocks, and the geofence events are all already being collected. The carriers are paying for the ELD device and the platform fee. The question is whether you have API access to that data, and the answer is usually yes for a small additional fee.

Geotab's MyGeotab API gives you per-vehicle GPS, ignition status, odometer, fault codes, and zone (geofence) events. The API is free to use if you have an authorized integration with the carrier's account. Samsara's Fleet API is comparable, with a per-API-key fee that lands around $10 to $25 per vehicle per month depending on the carrier's contract tier. Motive (formerly KeepTruckin) has a similar structure with an API tier that runs $15 to $20 per vehicle per month.

The carrier authorizes you, or the broker, to read their fleet data via a delegated API key. The data flow is direct. No middleware vendor in between. For most SMB brokers, somewhere between 60 and 85 percent of the carriers they regularly book with are on one of these three platforms. That covers the majority of your load volume already.

The 15 to 40 percent that is not covered is the owner-operator base and the smaller fleets running cheaper ELD options like Garmin, BlueStar, or various off-brand devices. For those carriers, you need a fallback method. The cheap and effective version is an SMS tracking link that pulls phone GPS with the driver's consent, sent at pickup and active until delivery. Not as clean as a real ELD feed, but it works for the loads where it has to.

What self-hosted freight tracking actually means

The phrase "self-hosted freight tracking" sounds like a lot of engineering, and it can be if you scope it badly. Done right, it is a focused stack with four parts, all running on cloud infrastructure you already pay for or can pay for cheaply.

The first part is the ingestion layer. A small service that polls the ELD APIs (Geotab, Samsara, Motive) for the carriers and vehicles you have authorization to track. The polling frequency depends on the lane and the customer's needs. Every 5 minutes is enough for most truckload moves. Every minute is overkill for almost everything and burns API quota fast.

The second part is the load record sync. Your TMS already has the pickup location, delivery location, appointment windows, carrier assignment, and customer reference numbers. The visibility stack needs to read those records, match them to the right vehicle in the ELD system, and stitch the GPS trail to the load. This is the integration that takes the most time to get right, especially if your TMS has a clunky API.

The third part is the ETA engine. Take the current location, calculate the remaining route to the delivery point, apply HOS-aware stopping logic, layer in a lane-specific multiplier from your own historical data, and produce a confidence interval. The first version of this is straightforward. The model gets meaningfully better after three to six months as you collect lane-level performance data.

The fourth part is the exception and alerting layer. Geofence the pickup and delivery. Fire alerts on pickup geofence exit, delivery geofence entry, dwell time exceedance, and ETA slip past the appointment window. Push to Slack, Teams, Twilio SMS, or email. A few hundred lines of logic and a notification fan-out.

None of these pieces are individually hard. The work is in stitching them together so they run reliably, recover gracefully when an API rate-limits you, and produce data your operations team actually trusts. That last part is the longest pole.

Real cost math for self-hosted on existing ELD APIs

Let us run the actual numbers on a self-hosted freight tracking stack for a broker doing 50 loads a day, with a carrier mix that maps to roughly 120 tracked vehicles per month across Geotab, Samsara, and Motive.

Build cost. A working stack with TMS integration, ELD ingestion for the three major providers, a basic ETA engine, a shipper-facing tracking portal, and exception alerting runs $22,000 to $38,000 with us as the build partner. The range depends mostly on which TMS you run and how clean its API is. McLeod and AscendTMS are on the lower end. Older systems with database-replication-only integration push to the higher end.

Ongoing hosting. A right-sized cloud setup on AWS or Render runs $180 to $400 a month for compute, Postgres with PostGIS, object storage, and notification infrastructure. The variance comes from how much historical data you keep hot (24 months of tracking data is dramatically larger than 90 days) and how many concurrent shipper portal users you support.

ELD API fees. Geotab is free at the API tier for authorized integrations. Samsara at 50 vehicles is roughly $750 to $1,250 a month. Motive at 40 vehicles is $600 to $800 a month. Call it $1,400 to $2,050 a month for 120 tracked vehicles across the mix. This cost scales linearly with your carrier base, and it is the largest variable cost in the stack.

Mapping and routing APIs. If you use HERE Maps or Mapbox for routing and traffic, budget $200 to $500 a month at this volume. OSRM self-hosted is an option that drops this to near zero in API fees but adds a server to maintain. Most brokers we work with start with HERE and move to self-hosted OSRM in year two if the volume justifies it.

Maintenance and operations. The honest line item. Either you have an internal technical owner or you have a maintenance contract with someone like us. A reasonable maintenance budget for a stack at this scale is $1,500 to $3,500 a month, covering API change adaptation, new carrier onboarding, ETA model tuning, and ops support. This is the number that gets undercounted in DIY math and is the most common reason self-hosted stacks fail.

Total monthly run cost lands between $3,300 and $6,500 for this size broker. Annualized, that is $40,000 to $78,000 a year in operating cost, on top of the one-time build.

Year-one total: $62,000 to $116,000 with build cost included. Year-two and three: $40,000 to $78,000 each. Three-year TCO: $142,000 to $272,000.

FourKites for the same broker over three years: $108,000 to $200,000 platform fees, plus $12,000 to $35,000 implementation. Three-year TCO: $120,000 to $235,000.

The numbers are closer than the marketing on either side would suggest. The self-host path wins clearly at the lower end of the broker volume range and gets squeezed at the higher end. The deciding factor is rarely the dollar amount. It is the operational ownership.

The carriers' ELD platform fee is doing a lot of work

The reason the self-hosted math works at all is that the carriers are already paying for the underlying telematics. A Samsara fleet at 20 trucks is paying Samsara somewhere around $30 to $50 per vehicle per month for the device, hardware, and platform access. That is $600 to $1,000 a month in revenue to Samsara that the broker never sees on their P&L.

What the broker pays Samsara is the API access layer on top of that. The carriers' base subscription covers the data collection, storage, and the carriers' own fleet management workflows. The API fee is for letting a third party (the broker, in this case) read the data with the carrier's consent.

FourKites, when it integrates with a Samsara-equipped carrier, pays the same kind of API fee. They just absorb it into their per-load platform price and you do not see the line item. The total cost of getting GPS off a Samsara-equipped truck is roughly the same whether FourKites handles it or you handle it yourself. The difference is who owns the integration, who owns the data, and who is responsible when something breaks.

Where FourKites actually wins, honestly

If you are seriously comparing FourKites against a self-hosted stack, the case for paying the FourKites bill is real for a specific set of brokers. Here is the honest version.

• Carrier network breadth. FourKites has direct integrations with thousands of carriers, including a long tail of smaller fleets that are not on Geotab, Samsara, or Motive. If your business is heavily weighted toward small carrier capacity, the FourKites coverage is genuinely valuable and hard to replicate.
• Shipper-named requirements. Some shippers, especially in CPG, automotive, and pharma, require FourKites as the visibility provider as a condition of awarding freight. If two or more of your top ten customers have this requirement, the discussion is mostly over before it starts.
• Multi-modal tracking. Ocean container tracking, rail visibility, intermodal moves, drayage with port system integration. FourKites has all of this built and operational. If you do meaningful volume in any of these modes, the build cost to replicate is much higher than for truckload-only operations.
• SOC 2 and vendor compliance. Some shipper procurement processes require a SOC 2 Type II report and a completed vendor security questionnaire before they will let you read their freight data. FourKites has these in hand. A self-hosted stack needs to clear the same bar separately, and that has its own cost.
• Predictive ETA at the top end. FourKites has years of training data across hundreds of millions of shipments. Their predictive ETA in the 95th percentile of accuracy is genuinely better than what you can build internally without serious investment. At very high load volumes, this gap matters operationally.

If three or more of those describe your business, stop reading and call FourKites back. The self-hosted path is the wrong answer for you, and we will tell you the same thing if you ask us directly.

Where self-hosted freight tracking actually wins

The opposite case is also worth being explicit about. Self-hosted is the right answer when the following are true.

• Your carrier base is concentrated in mid-size fleets running Geotab, Samsara, or Motive. Above 70 percent coverage from those three providers is the sweet spot.
• You move domestic truckload and reefer, with limited or no multi-modal exposure.
• Your shipper customers want tracking but do not name a specific platform. They care about a working map and accurate ETAs. They do not care whose logo is on the portal.
• You have, or can hire, or can retain via a maintenance contract, a technical owner who can operate the stack. Not just build it. Operate it at 11 p.m. when Samsara changes their auth flow.
• Your per-load margin is tight enough that a $2 to $5 per load platform fee meaningfully affects your operating profit at year end.
• You value owning your tracking data for downstream analytics, customer-specific reporting, or lane-level pricing intelligence, without going through a vendor's product roadmap to get features built.

If four or more of those are true, the self-hosted path is the lower-cost, higher-ownership option, and the build math pays back inside 18 months for most brokers in the 30 to 100 loads a day range.

The decision framework at three volumes

15 to 30 loads a day. FourKites is not really available to you at a price that makes sense, and the self-host build cost is hard to amortize at this volume. A middle-tier paid tool like Trucker Tools at $1,500 to $3,500 a month is usually the right answer. Revisit the build path when you cross 40 loads a day or when your carrier mix consolidates around ELD-equipped fleets.

30 to 100 loads a day. This is the self-hosted sweet spot. The build cost amortizes over enough loads to pay back in 14 to 22 months. The operational burden is real but manageable with a partner relationship. The freight visibility cost on a per-load basis lands at $1.50 to $3 once the build is paid back, versus $2 to $5 on FourKites. The difference adds up to meaningful operating profit at this volume.

100 to 200 loads a day. The decision gets harder. FourKites at this volume drops to $0.80 to $1.50 per load after negotiation, and the carrier network breadth becomes more valuable as you book a more diverse capacity mix. The self-hosted stack still works but needs a dedicated operations function, not just a maintenance contract. We have brokers in this range running self-hosted successfully, and brokers in this range who tried it and ended up on FourKites two years later because their internal IT person left and they did not replace the role.

200+ loads a day. The platform fee gets diluted enough that FourKites is rarely the worst answer. Negotiate hard on the multi-year contract, use the self-host option as a real BATNA, and you can usually land a per-load rate that makes the platform fee worth paying. The brokers we see at this scale who still self-host are the ones with strong internal engineering teams who treat the visibility stack as a competitive advantage, not a cost center.

What the migration actually looks like

If you decide the self-hosted path is right, the migration timeline depends on where you start. From spreadsheets and check calls to a working stack runs 7 to 11 weeks end to end. The longest pole is TMS integration, especially if your TMS is older. Plan for two to three weeks of dual operations after go-live while ops gets comfortable.

From a paid tool like Trucker Tools or MacroPoint, the migration is 8 to 12 weeks. The carriers already have a tracking workflow they are used to, which makes the carrier-side conversation easier. The shipper-side cutover needs a structured comms plan, especially if shippers are used to logging into a specific portal.

From a FourKites pilot or active contract back to self-hosted, the migration is realistic but needs planning three to four months before the contract renewal date. Most of the value is reproducible. The piece that takes the longest is rebuilding the long-tail carrier coverage that FourKites gave you "for free" as part of their network. Plan for some loads to drop off the visibility map during the transition and a comms plan for shippers.

The reverse migration, from self-hosted to FourKites, is the easiest of all. The data flows already exist. The shipper portal can be migrated. The integration work has been scoped. Brokers who outgrow the self-hosted stack typically make this transition cleanly in 6 to 10 weeks.

The questions to answer before you decide

When a broker asks us whether to sign with FourKites or build a self-hosted alternative, the answer comes down to four questions.

First, what percentage of your regular carrier base runs Geotab, Samsara, or Motive? Pull the list, sort by load volume, and tag each carrier. If the top 70 percent of your load volume is on those three platforms, the self-hosted path has a real foundation. Below 50 percent, the math gets harder fast.

Second, do any of your top ten shipper customers explicitly require FourKites or another named visibility platform? If yes, that customer alone may justify the FourKites bill. If no, you have a real choice.

Third, do you have a technical operator for the stack? Not a builder. An operator. Somebody who answers the phone when an API breaks at 3 a.m. or wakes up the next morning and fixes it before the operations team notices. If the answer is no, and you are not going to hire one or contract one, the self-hosted path is not for you regardless of what the cost math says.

Fourth, what does your 24-month load volume projection look like? Below 30 loads a day, stay on a paid tool. Above 75 loads a day, the self-hosted path pays for itself comfortably. In between is where the rest of the factors decide it.

The FourKites alternative conversation is not a question of whether FourKites is good. It is. The question is whether your business is the customer FourKites was built for. For a meaningful portion of the SMB freight broker market, the answer is no, and a self-hosted freight tracking stack built on the ELD APIs you and your carriers already use gets you 80 to 90 percent of the value at 50 to 70 percent of the three-year cost. The other 10 to 20 percent is real, and being clear-eyed about it is how you make this call without regretting it in year two.