Can your Passenger Wi-Fi network pay for itself?

Connections = Ridership / Visitors * Take rate

Take rate depends on availability and prices of cellular data, as well as Wi-Fi Quality of Experience (QoE). The latter is under operator’s control, however there are certain physical limitations: for instance, if you use LTE backhaul, you should consider coverage and available bandwidth while keeping in mind that users’ personal data package is usually available with somewhat the same quality

Our studies generally show that, unnecessity of free Wi-Fi due to decent cellular coverage accounts for 40% of unreached target audience while QoE affects around 30%. The remaining 30% can be attributed t.

If you do not have Wi-Fi, you can apply the following sample take rate figures to your preliminary business model (and adjust based on pilot roll-out actual figures) for free Wi-Fi with a simple captive portal logon:

• Cafe/restaurant — 3-6%
• City over ground transport, including buses, trams, trolleys or light rail — 3-7%
• Shopping malls — 5-10%
• Subway stations — 5-15%
• Stadiums — 10-20%
• Airports — 10-30%
• Subway trains — 20-60%
• Long-distance railways, aircrafts and marine cruise liners — 20-50%.

We would recommend to use both figures from the suggested range as a worst-case and a best-case scenario.

But if you want to show more than one commercial in a row to work with the next parameter, Impressions, you will obviously affect the Quality of Experience and therefore drive usage decline. It’s especially the case for venues with repetitive visits pattern, when user base consists of a stable core (city transportation in particular).

The effect of adding advertising impressions is not linear at all and sometimes the fifth banner will lead to a massive decline while the fourth is absolutely fine. However, this can be estimated quite accurately with a help of peer reviews.

Generally, if you don’t have specific venue requirements to advertising, we would suggest the following standard ad sequence:

• Cafe/restaurants, city over ground transport — sponsored welcome page + one interstitial / video
• Shopping malls, Subway stations and Stadiums — sponsored welcome page + one or two interstitial / videos
• Airports — sponsored welcome page + two interstitial / videos
• Subway trains — sponsored welcome page + three or four interstitial / video (limited to 60 seconds overall)
• Long-distance railways, aircrafts and marine cruise liners — 20-50% — sponsored welcome page + three or four interstitial/ video (limited to 120 seconds overall).

However, the latter is quite specific — if you use satellite backhaul, it’s unlikely possible to reach decent Revenue/ OPEX ratio.

Even when implementing these ad sequence recommendations, for repetitive patterns you can expect up to 50% decline in usage and for non-repetitive

eCPM stands for «effective Cost per Mille». To calculate eCPM you should divide your overall ad revenue by the amount of impressions you could generate with given amount of connections, even if some portion is unsold.

In Wi-Fi networks, eCPM varies from nothing to hundreds of dollars, depending on country and market conditions, quality of your sales, targeting capabilities and technologies networks apply, amount of ad inventory for sale, etc.

The more original network’s engagement mechanics and pricing models are and the less inventory it generates (there is some entry barrier, however) — the more likely it’s possible to get unfairly high prices. But this effect dilutes once the network reaches significant inventory volumes. The reason is simple — by reaching that point, the network becomes an advertising market player, and as any other market it has its rules. In most cases, advertiser who buys 1 million impressions at least will compare cost per reach (CPR) across various media channels. And Wi-Fi ads, as a premium advertising product, will require proper skills to sale it at scale.

The main consideration here is that Wi-Fi advertising is much more efficient than ordinary internet advertising with a mobile audience of 100% mobile audience, and if suitable sales team and technology are in place, it’s possible to get at least 2-3 times higher eCPM than average for premium mobile internet resources in a given area.

Okay, we are all set. Let’s apply the above to some abstract Subway Stations Wi-Fi, where stations are well covered with cell networks and average 1Gb mobile data package prices are moderate.

Ridership is 2M/day, take rate 10%, branded authorization plus one full-screen interstitial banner at $5 eCPM. Daily revenue will be 2M*10%*($5/1000*2) = $2,000.

We are not trying to cover cost side considerations in this article. In our experience, squalityand high-density Wi-Fi networks will not pay back with advertising. But there are some that can. Our main area of interest is on-train Wi-Fi; thus, we used a simple Ridership Density multiple (e.g. daily ridership per kilometer of track, which should be replaced with Connections per Dollar invested if you have accurate data for calculation) and compared all subways around the world. Good ridership density is around 20K, minimal to try pure advertising model is ~14K. Also, there is scale threshold, because the smaller the system, the more expensive 1 km of infrastructure. Apparently, only 20 lucky subway systems in the world have more than 1.3M ridership with more than 14K ridership density..

Other subways would have to consider additional models to launch public Wi-Fi