Before judging a campaign, use attempted sends, provider acceptance, bounces, delivered events where available, and any unresolved or unobservable outcomes. These counts tell a manager whether the audience had a reasonable chance to receive the message. Opens, clicks, and orders answer different questions.
Start with the person responsible for the campaign, usually a marketing lead or manager. That person should reconcile the report before a shift leader acts on it. A lunch offer sent an hour before service needs a quicker check than a newsletter planned for next week. An unresolved delivery count may be tolerable while a send is still processing, but not when the offer has already expired.
A restaurant marketing campaign delivery report is useful only if its labels have clear meanings. Ask each provider what it can observe, when the numbers update, and whether the report separates failures from events it cannot see.
Choose the decision before reading the report
The first decision is whether to judge the campaign now, wait for the report to settle, or investigate a delivery problem. Do not treat an accepted send as proof that a guest received it. Acceptance generally means a provider or downstream service took responsibility for processing the message; the exact handoff depends on the channel and provider.
Likewise, a delivered event is not proof that a guest read the message, visited the restaurant, or used an offer. Even when a report shows a delivered event, ask what system recorded it and what it means for email or SMS. A provider may have different definitions for each.
The manager should also check the audience before looking at results. Was the intended list selected? Were duplicate contacts, invalid addresses, and people who should not receive marketing handled as intended? Access to a guest's contact information is not the same as permission to market to that guest. Assign someone to review the list and the restaurant's permission rules before sending.
Service conditions change the decision. If a short-notice special has expired, a late delivery report cannot restore the missed opportunity; it can still help staff decide whether to change the next send. If a message concerns a future event, the manager may have time to investigate and make a separate, carefully targeted follow-up. Avoid automatically resending to everyone: some guests may already have received the first message.
Build a delivery funnel with an owner at each step
Use one channel and one campaign at a time. Record the audience selected, the number attempted, and the time at which the report was checked. Then map provider statuses into mutually exclusive groups so each attempted message is counted once. Keep the provider's original labels alongside your own worksheet categories.
- Audience review: The campaign owner confirms the intended segment and records any contacts excluded before an attempt. Exclusions should not be added to attempted sends.
- Attempt and acceptance: The campaign owner records attempts and separates messages accepted for processing from those rejected before acceptance. Ask whether retries are counted as new attempts or attached to the original message.
- Delivery outcome: The report reviewer divides accepted messages into delivered, bounced or failed, pending, and unknown outcomes, using the provider's definitions. A bounce or failure is a reported problem, not an estimate of how many guests ignored a message.
- Action: The manager decides whether to wait, investigate, or review campaign response. The staff member assigned to the issue records the next check time and any manual work.
For a hypothetical email campaign, suppose the owner selects 124 eligible contacts. Four are excluded before sending, leaving 120 attempted messages. The provider accepts 114 and rejects six. At the review time, its accepted messages have mutually exclusive outcomes: 100 delivered, eight bounced, four pending, and two unknown.
The arithmetic balances: 114 accepted plus six rejected equals 120 attempted. Within the accepted group, 100 delivered plus eight bounced plus four pending plus two unknown equals 114. The four pre-send exclusions are outside that funnel. If the provider reports bounces at a different stage, the owner should adapt the worksheet rather than force its labels into these categories.
The manager's decision is to investigate the six rejections and eight bounces, then check the four pending outcomes later. The two unknown outcomes stay labeled unknown. The manager can review response from the reported delivered group, but should not call the campaign a poor performer based on all 120 attempts as though every guest received it.
Compare reporting approaches before buying
The table compares ways to run the check, not verified features of particular vendors. Use it to decide how much staff work you can accept and which reporting details belong in a demonstration or quote.
| Approach | Staff effort | Guest impact | Cost and vendor questions |
|---|---|---|---|
| Summary counts in the campaign screen | Low if statuses are clear; higher if counts do not reconcile | Broad totals may not identify who should be excluded from a follow-up | Ask which events are included, when they update, and whether reporting costs extra. |
| Message-level report or export | More review and secure handling of contact data | Can help staff avoid sending a follow-up to the wrong people, if outcomes are identifiable | Ask about export access, retention, channel coverage, and any usage or setup charges. |
| Manual log using available provider records | Highest ongoing effort; a named owner must reconcile each send | Slow checks can delay a correction or follow-up | Ask whether the records needed for the log are accessible and what staff time the process requires. |
A summary may be enough for a restaurant that sends an occasional newsletter and can wait for totals to settle. A restaurant running time-sensitive offers may value message-level outcomes and a clear update schedule more highly. Neither approach makes an unobservable event observable.
Compare written quotes on the same basis: recurring charges, setup, included message volume, per-message charges, reporting or export fees, and staff time. Ask how failed attempts and retries are charged. A low quoted message rate alone is not the full operating cost. If a quote leaves an item open, record it as unknown rather than treating it as zero.
Fleksa describes email and SMS campaigns and segmentation in its marketing tools. Ask to see the delivery events, exports, and send controls available for your setup rather than assuming a particular report is included. Its campaigns documentation can help frame that conversation. If an offer is tied to loyalty, review delivery first and reward use separately; receipt of a message does not establish that a guest earned or redeemed a reward.
Run a reconciliation and exception test
Before selecting a provider, ask for a demonstration using simulated records or a sandbox. Practice must not send a real guest message, charge a card, or dispatch a courier. Provide the vendor with sample statuses and ask staff to complete the worksheet themselves. Then, if appropriate, plan a separate live service pilot with approved recipients and agreed send controls.
Pass the count test when the reviewer can account for every attempt exactly once at each defined stage, explain exclusions, and identify the report's cutoff time. The sample above passes that arithmetic test. It does not prove every message was delivered: pending and unknown remain separate outcomes.
Fail the count test when totals cannot be reconciled, a status definition is unclear, or retries might be counted twice and nobody can resolve the difference. Record the screen or export used, the campaign identifier, channel, time checked, status definitions, and the vendor's explanation. Do not fill a gap by calling it delivered or failed.
Test exceptions individually. Ask how the report represents an invalid destination, a rejection before acceptance, a bounce after acceptance, a delayed result, and a status the provider cannot observe. Expected behavior is a distinct label or a documented way to account for each case. Also ask whether a later update changes the original report and whether a guest-level record remains available for follow-up decisions.
If a vendor cannot show an event, mark it not observable in the worksheet. That differs from pending, where a result may still arrive, and from a confirmed failure. The practical fallback is a manual log of available totals, a scheduled recheck, and a decision rule that prevents a broad resend while outcomes are uncertain. If message-level evidence is unavailable, staff may need to avoid a targeted resend altogether rather than guess who missed the first message.
For the next buying conversation, bring one recent campaign's anonymized counts and ask for the same reconciliation in a demonstration of Fleksa restaurant marketing automation. Compare the result with your current process, then request a written quote and confirm which capabilities apply to your setup.
Frequently asked questions
Is an accepted message the same as a delivered message?
No. Acceptance indicates a processing handoff under the provider's definition. A delivered event, when available, is a separate reported outcome. Ask what each label means for your channel.
Should a manager judge a campaign while messages are pending?
Usually, wait for the expected reporting window if the decision can wait. If the offer is time-sensitive, document the pending count and make a limited operational decision without treating those messages as delivered or failed.
Can a delivery report show whether a campaign generated orders?
Delivery reporting checks message outcomes. Orders and campaign attribution require separate measurement. Start by establishing which recipients may have received the message before interpreting response.
