A tender alert is a saved search that runs automatically against newly published notices and notifies you when something matches. You define the filters — usually CPV codes, countries, value range, and keywords — and the system emails you or flags the opportunity in your dashboard. Alerts solve discovery. They do not solve deadlines: for that you need a separate routine that starts the moment an alert arrives.
The two problems get confused constantly, and it is expensive. Most lost opportunities are not lost because nobody found the tender. They are lost because it was found with nine days left, sat in an inbox for four of them, and the qualification paperwork took longer than the three days that remained.
How an alert is actually built
Every alert has four moving parts, and each one is a place where opportunities get silently dropped.
The filters
CPV codes should be the backbone, because they are assigned by the buyer and do not depend on how the notice was worded or which language it was written in. Around 97% of active tenders in the Tendly snapshot carry one. Keywords are a useful supplement but a poor primary filter — they miss anything phrased unexpectedly and they break entirely across languages. Location and value range then trim the result to what you can realistically deliver.
The source
An alert can only fire on notices its source can see. A national portal's own alert service covers that portal only; a supranational journal covers above-threshold notices only. If you rely on one source you are, by construction, blind to the rest. This is the strongest argument for alerts built on aggregated data rather than portal-by-portal subscriptions.
The frequency
Daily is the right default for most businesses. Real-time is worth it only in categories where deadlines are habitually short. Weekly digests are comfortable and quietly dangerous: more than a quarter of all active tenders are closing within seven days, so a weekly cycle can deliver an opportunity that has already expired.
The ranking
A filter-only alert returns everything that matches, which for a broad CPV division can be dozens of notices a day — and an alert people stop opening is worse than no alert. Ranked alerts put the best-fitting tenders at the top with a reason attached, which is what AI matching adds on top of the filter.
Why deadlines get missed anyway
Four failure modes account for almost all of it:
- The alert arrives late in the tender's life. The clock started when the notice was published, not when you saw it.
- Nobody owns the alert. An alert sent to a shared inbox belongs to nobody. Assign a named owner who triages within one working day.
- The qualification pack is not ready. Tax clearances, insurance certificates, and audited accounts take days to obtain — days you do not have. Keep the standing document pack current so a bid decision is never blocked on paperwork.
- The deadline moved and nobody noticed. Amendments and clarification rounds change dates in both directions. Monitor the tenders you are working on, not just the ones you are looking for.
A deadline routine that survives a busy month
Work backwards from the submission deadline rather than forwards from today. A simple T-minus schedule, adapted to the length of the window, prevents almost all avoidable losses:
- Day 0 — triage. Within one working day of the alert, decide pursue, watch, or discard. Discarding fast is a feature, not a failure.
- Day 1 — download and read the pack. The notice is a summary; the mandatory requirements are in the documents. Build the compliance checklist here.
- The clarification deadline is the real first deadline. It falls well before submission and it is the last moment you can ask the buyer a question. Diary it separately — it is the deadline people forget exists.
- T-minus one week — content complete. Everything drafted, every mandatory document attached, pricing agreed internally.
- T-minus two days — submit. Not on the final morning. Portals throttle under load, uploads fail, files exceed size limits, and signatures need chasing. Submitting early costs nothing and removes an entire class of risk.
If it helps, run the readiness check before you commit rather than after: the tender readiness checker asks the questions an evaluator will ask about your eligibility, and the bid/no-bid helper structures the triage decision.
Alerts for contracts that have already been awarded
One category of alert is consistently underused. Award notices — published when a contract is signed — tell you who won, at what price, and when the contract ends. That is the raw material for two useful plays: approaching the winner about subcontracting, and diarising the re-tender date so you are prepared months before the next competition opens rather than reacting to it. Set up an alert on award notices in your categories alongside the one for open opportunities, and read what an award notice contains if the format is unfamiliar.
Tuning alerts so you keep reading them
Alert fatigue is the main reason alert systems fail, and it is a tuning problem rather than a volume problem. If you are discarding most of what arrives, your filters are too broad — narrow the value range or drop a CPV division that keeps producing work you would never take. If you keep discovering relevant tenders by other means, they are too narrow: add adjacent codes, widen the geography, or check whether the tender was classified under a code you had not considered. Building the CPV list properly is usually the fix.
Review the filters quarterly, and always after you win or lose something unexpected. What you bid for changes faster than most alert configurations do.
Put the routine somewhere it will be followed
The best alert configuration is worthless if the follow-through lives in someone's head. Give every pursued tender an owner, a deadline in a shared calendar, and a compliance checklist someone signs off. Then let the software do the finding — you can browse what is currently open by country and category to see what your filters would be catching today.