Start with your traffic, not with a feature list
Write down who sends to you and how (SMPP binds, HTTP integrations, a portal), who you deliver to (SMPP suppliers, HTTP suppliers, SS7), your monthly volume and peaks, and how you bill (prepaid, credit, by submitted or by delivered). Every vendor demo should be measured against that page.
Ask for an evaluation on your own routes and hardware. Throughput figures without your connections and configuration tell you little.
Customer and supplier connections
An SMPP server is the front door. Check that customers can bind over TLS, that throughput is limited per credential and per account, and that the platform acknowledges a message only after it has stored it. Ask what happens to a multipart message when a part is retransmitted: it should be neither sent nor billed twice.
On the supplier side, look for SMPP 3.4 and 5.0 binds with TLS, per-trunk throughput and uptime history, and the ability to change trunk settings without restarting. If you terminate on SS7, ask exactly how: over SIGTRAN, which MAP operations are supported, and who manages global titles, point codes and links. See SMPP server software.
Routing you can explain
Routing decides your margin and your delivery rate. You should be able to read why a message took a route, test a change before it touches traffic, and pick suppliers by cost, weight, split or quality. Supplier errors should map to clear actions: retransmit later, reroute, or stop. Margin and quota guards prevent the expensive mistakes. See SMS routing and LCR.
Billing that closes on the day’s numbers
Ask when a message is charged and how the platform guarantees it is charged once. Rating per segment, balances with credit limits, supplier price imports and invoices built from the same records as traffic remove the month-end reconciliation. See SMS billing software.
Traceability for support and disputes
When a customer says a message never arrived, your team needs the pipeline with timings, the supplier response and the raw PDUs, for that single message. Ask the vendor to open one live in the demo.
Security and control
Look for IP allowlists per credential and per API key, automatic blocking after failed binds, hashed API keys, permission-based roles and an audit log. Number lookup and an SMS firewall protect both your network and your margin. If the platform offers AI assistance, ask what data leaves the platform. See firewall and HLR and the AI console.
Deployment and ownership
Decide whether you need the platform on your own servers or a dedicated cloud installation, and whether it must carry your brand for your customers. Ask how updates and database migrations are delivered. See on-premise and cloud and the customer portal.
The checklist
Tick what your shortlisted platform covers. Anything left unticked is a question for the next demo.
