How to Audit Your Mobile Advertising Performance and Stop Budget Bleed

Many restaurants which run ads during the dead hours between 2 pm and 5 pm have a serious leak in their advertising spend. Everything looks good, a fair click-through rate, enough impressions – but no one knows whether these clicks turned into actual customers. An audit will take about an hour, and you will find several issues to resolve right away to reallocate the money and fund the next month’s campaign better.

There is no need to change platforms or agencies; use the current data, answer a few questions that an entrepreneur might not have known were worth asking.

First, pull 90 days of data before any judgment.

The campaign’s general numbers are pretty misleading – and a three-times return on ad spend (ROAS) hides the fact that the campaign is driving traffic somewhere it should not. When the data is segmented by the campaign, ad set, and placement, you can see which of the underperforming ad sets and placements is dragging the whole campaign down. A six-times ROAS is significantly better than a zero-point-four-times ROAS, so which of the two you would like to shut down and reinvest the money into the winner?

Next, check if the tracking is working correctly.

It might seem too obvious, but the main reason many campaigns are paused prematurely is that the advertising tracking is broken. If the pixel does not fire when you click on the call now button, get directions, or land on the mobile site, every campaign looks terrible even when it delivers customers. The simplest way to check if everything works is to manually click on the ad and see whether the conversion event appears in your dashboard in a matter of seconds. The same goes for other events – get directions, call now, location visit. A restaurant that tracks only reservation conversions misses out on half of the funnel, and the other half has nothing to do with reserving a table right now – therefore, it does not make sense to blame the ROAS when the whole funnel is ignored.

Then, scan the data for invalid traffic patterns.

Invalid traffic patterns are those where the clicks come from irrelevant users or bots. The obvious signs are high click rates with low session duration and random traffic spikes on a daily level or from one IP address in particular. Invalid traffic patterns, however, often have nothing to do with bots; they are signs that a campaign’s conditions and placements target the wrong users. For example, in-app advertisements have much higher risks of invalid traffic since some apps are notorious for having multiple mediation layers that generate ad impressions invisible to the users.

To find invalid traffic patterns, take a click-through data report and sort it by session duration. Take a closer look at any traffic source with average session duration below three seconds and no downstream page views; these are the ones to exclude from further consideration. If you see traffic spikes from one IP address or an entire device class, these are also signs of bots. Once you remove such resources, you will see a jump in conversions in the sources that remain on the list.

Match dayparting to the actual quiet hours.

An audit of such a restaurant campaign becomes much more specific from this point on. Take the dayparting report and put it alongside the actual hourly reservation and walk-in numbers – ideally, an ad campaign analytics tool that has access to both would be beneficial at this point. The main finding here should be the hours that generate impressions and clicks but have zero reservations or walk-ins.

Many restaurants set the dayparting to the same schedule, based on their best guesses as to when they are the busiest. Static geo-fencing adds to the problem since the same ads appear to everyone within a three-mile radius, regardless of whether that user has a reasonable chance to visit the restaurant before the lunch period ends. Dynamic dayparting based on a much smaller radius helps keep the ads relevant to the users who are close enough to the restaurant to be enticed by a special offer.

Similarly, dayparts that were scheduled around a 9 pm to close timeframe often have decent conversion rates, but the 3 pm slot regularly receives traffic with no conversions – yet many restaurants ignore the former and keep the latter simply because their original schedule was based on assumptions.

The key is finding when the restaurant ads turn impressions into reservations during the dead hours specifically. Impression volume during the busy hours is not important – but impressions turned into reservations during the dead hours will help both the restaurant and the advertising network.

Tighten the conversion window.

Most ad platforms use a thirty-day click conversion window as a default, but it might be too long for restaurants. Very few users will see the ad on Day 1 and reserve a table on Day 30, so a seven-day conversion window is a more honest representation of the campaign’s impact on the restaurant’s performance. The ROAS will reduce significantly, but after the next seventy-two hours, the restaurant owner will understand that the previous numbers were artificially pumped up.

Moreover, a one-week click conversion window will show which ads influence the users’ decision to visit the restaurant on the same or following day better than the thirty-day view. This way, the restaurant will have a more accurate insight into the impact of its advertising spend.

Rotate the creatives before they become irrelevant.

The longer the ad is in rotation, the lower its click-through rate (CTR) is likely to be. As the same users see the same creatives over and over, the CTR drops, but the restaurant’s budget does not reduce – so the temptation to raise the bid for the next rotation is inevitable, and the results will be no better than with the previous ad. Instead, the restaurant should decide on a CTR benchmark and rotate the creatives when the CTR is lower than this number. A good rule of thumb is to use a CTR benchmark of approximately 0.5% for any creative that has already been in rotation and thus lost its freshness.

This tip is particularly useful for restaurants since their creative assets tend to be few and far between – several food pictures and a logo. Try using different dishes, different calls to action, and different hours in the same ad set. A small change to the CTR could be enough to make the campaign more effective without spending an additional cent on advertising.

Remove placements that do not deliver conversions.

Placement-level analysis is rarely done during audits since it is time-consuming, but it is one of those places where a small change can yield considerable results. Run an impression-level report to check which apps or websites consume the most budget with no conversions. There is a good chance that you will find several placements with an unreasonable share of impressions and no visitors, directions, or phone calls.

These placements are not necessarily fraudulent or low-quality – they simply became a waste of your advertising budget. Exclude them from the campaign to reallocate the impressions towards the placements that actually converted. Removing a single low-performing placement could be more beneficial than a few months of meticulous spend optimization.

Reallocate spend based on CPA per daypart and geo-zone

After the necessary changes have been made to the campaign, it is time to address the budget directly – and this part is actually very simple. Set the CPA benchmark for each daypart and geo-zone, and reinvest the overperforming budget into the underperforming ones to meet the CPA level. For example, if one geo-zone delivers eight-times ROAS and another only delivers one-times ROAS on a similar budget, you are essentially using the former to subsidize the latter due to a uniform daypart spend.

A similar CPA-based approach applies to dayparts – if one daypart has a significantly lower CPA than the others while maintaining a decent ROAS, you should reinvest the extra budget into the dayparts that need more impressions. This step completes the audit since it actually reallocates the budget based on the findings.

Benchmark your network against an alternate supply.

Your network’s fill rates and CPMs are difficult to evaluate if you have nothing to compare them to. Launch a test campaign, or enough impressions to collect meaningful data, with another supply source and compare the two networks’ CPM, viewability, and invalid traffic levels. A good audit will also help you determine whether your demand-side platform (DSP) meets the requirements for the best mobile ad network for your restaurant, including fraud protection, viewability, and fill rates. Sometimes the original supplier is better than the new one, and you know now that you chose wisely. Other times, you have found a new ally to help optimize your restaurant’s advertising campaign. Either way, you now have a reliable benchmark to measure future performance against.

Turn this into a 30-minute monthly audit.

The first audit took most of the afternoon, but subsequent ones should only take about an hour. Set a short reminder every month with the above checklist – check tracking, look for invalid traffic patterns, analyze dayparting in depth, observe the CTR, remove low-performing placements, and reinvest the extra budget. The faster you identify budget leaks, the fewer wasted impressions you will see – and the less you will have to convince yourself that mobile advertising is not worth the money spent on it.

A few hundred dollars’ worth of wasted impressions every quarter adds up – but a few thousand wasted impressions only make the restaurant owner wonder whether mobile advertising is really worth it at all.