How can travel startups hang on these days given the dominance of a few incumbents?

It’s hard to be optimistic about the chances of consumer-facing travel startups to succeed when Expedia Inc., the Priceline Group, and Ctrip appear to have the muscle to crowd out new entrants.

These questions come to the fore as Hopper, a mobile-only travel startup that has raised $78 million, begins to expand beyond airfare predictions and flight bookings to offer consumers hotel rate predictions and room bookings — beginning with New York City.

Hopper will have to compete with the incumbents, which enjoy so-called “network effects.” By attracting more consumers and spending collectively about $9 billion in annually on marketing, they create attractive markets for hotels and airlines. And by attracting more suppliers, they make their marketplaces more appealing to consumers — which, in turn, creates a cycle that makes the platforms grow.

Even startups backed by major companies, like Alibaba’s Fliggy and Ctrip-supported MakeMyTrip struggle to gain traction quickly. One problem: The rising cost of customer acquisition because of gatekeepers such as Google and Facebook.

Despite being founded in 2007 and executing a series of missteps, Hopper only found its footing in the past two years after it become a mobile-only business focused on airfare predictions.

“We see that mobile is different because our apps are holding our own against the big guys,” said Hopper founder and CEO Frederic Lalonde., pointing to data from App Annie, a research firm, that Hopper’s app has remained for months one of the most downloaded travel apps in multiple countries. Last month in the U.S., only Booking.com had more app downloads than Hopper in the travel category.

Since the company launched its apps, consumers have downloaded them 17 million times, the startup said.

Hopper markets its apps on social channels like Pinterest and Instagram.

Hopper, which acts as a travel agency and is the merchant of record on transactions, said it is on track to process $500 million in gross bookings in 2017. That’s roughly triple the pace of the year before, the company said, but it is only a sliver of the $72 billion that Expedia processed in gross bookings last year. Gross bookings are the full value of bookings, a much greater amount than the revenue that comes to the seller.

Like most private companies, Hopper wouldn’t disclose its revenue. But it said its users are loyal, and that a majority of its revenue is now from repeat customers.

Hopper’s strides point to the possibility that network effects on desktop don’t necessarily port to mobile. In other categories, disruptors have had success via mobile. Duolingo has outmaneuvered established language-learning brand Rosetta Stone, bill-payment tool Venmo has snuck up on major bank brands, and Tinder is overtaking Match.com — to name a few examples.

While Hopper appears to have carved out a small niche market — millennial users who will be loyal to a superior mobile experience and the novelty of flight-prediction tools — it faces a challenge as it tries to expand geographically. Will it be enough to become the acquisition target of a company like Uber or Amazon if Hopper fails to expand beyond its base of frequent travelers who care about the efficiency of a mobile user experience and the novelty of price forecasting? Can Hopper dream of an initial public offering one day?

Transaction volumes on mobile remain low, with perhaps fewer than 5 percent of all online travel booked by mobile, according to some industry estimates. As the volumes grow, the big online players may turn their big guns on the sector that Hopper participates in.

Today’s openings for new entrants may narrow over time. Already, Booking.com has been tip-toeing past its hotel-only focus to recently add to its homepage features tabs for flights from Kayak, car rentals from RentalCars.com, and restaurant reservations from OpenTable.

As Hopper expands into hotels, it also starts competing with HotelTonight, a mobile-first travel company that has tracked a similar path — surviving and becoming profitable but remaining a relatively small operation. This autumn, HotelTonight expanded its booking window to up to 100 days in advance of arrival, nearly identical territory to what Hopper is hoping to enter.

Several industry observers are skeptical that HotelTonight can grow enough to have an initial public offering without having truly distinct inventory from what the giant players have.

There are similar sentiments about Hopper, which also intends to only display a handful of hotels for any given date in only a selection of popular destinations — not unlike HotelTonight’s model. But Hopper CEO Lalonde said that consumers book flights on its app an average of 140 days in advance of departure. About a fifth of purchases are driven by recommendations made by artificial intelligence-based suggestion software, the company claimed.

Hopper has a hedge against one problem facing travel startups — that people only interact with the business and travel once or twice a year — because its flight predictions involve millions of push notifications to its user base, keeping consumers more engaged. Hopper claimed its engaged users interact with the business four times a week on average.

The company is hoping to further increase monthly engagement by adding hotel bookings and rate predictions. Flight itineraries for tickets booked through the app average eight days, so potentially there is a lot of incremental hotel spending to attach to airplane ticket sales.

The company is in the process of opening up locations abroad in its expansion effort. A number of companies, including Farecast, DealAngel, and Kayak have previously dabbled in the type of price predictions that Hopper is involved with. Hopper will making predictions on how individual property rates may go up or down.

Hopper skeptics abound, especially given the company’s decade-long saga to get to this point, but the 100-person company is expressing confidence in its near-term future. It is hiring for a few dozen positions in software engineering, and for establishing direct ties with hotels.

Photo Credit: Founded in 2007, Hopper initially saw itself as building a database of flight and destination information and considered options as a consumer brand or a business-to-business offering. Now it is solidly a consumer app. Hopper