How accurate is an airDND report?

Question

How accurate is an airDND report? I have an Airbnb report run on a property that I was considering fir a short term rental.  Once I backed out mortgage, taxes, insurance, lawn care Internet,  etc it didn’t really show much more than a breakeven profit if I was to use a property manager. The property manager I spoke with seems to be very knowledgeable about short term rentals and was saying that the report is usually low. I’m not sure if he means it’s low because he does a better job at filling the properties.

Michael Baum · Answer

Hey @Joe S., I play a Dwarf landlord in AirDND. I have a +2 to maintenance and roll a D6+2 for evictions. Luckily I have a pretty high charisma for a dwarf so I can usually talk people into paying their rent on time. :)

Seriously though, AirDNA can be real hit or miss. The more input the more accurate it will be.

What @Justin Anderson said should be a red flag for you. PF/GB is a very STR dense area with lots of data inputs. If they are off on their calculations for nightly rates, then you can bet it will be off elsewhere.

In my area, AirDNA radically overestimated my nightly rate. If I had invested based on that, I would be bankrupt. What I did was pull comps in off AirBNB and VRBO that had homes similar to mine, plugged it in to a sheet and came up with my nightly rate.

Do you have to use  PM? You will find that most of your investments will not be profitable if you do.

Paul Sandhu · Answer

It's 100% accurate.  I checked my own market and the numbers I got were directly in line with my own rent numbers.  Then I dug a little deeper.  My rental property that I had listed on Airbnb was the only property that was used in their calculations.  So of course, with a sample size N = 1, it's going to be 100% accurate.

Justin Anderson · Answer

I find it not so accurate in Pigeon Forge/Gatlinburg.  For example it has my personal cabins making about 30% less than they do for the small ones and 50% less on the big ones.You have to remember that it shares data on the area for what the typical property would do.  That means 50% of properties will probably do better.  Will you run your place better or worse than the typical property manager?  I find it very easy to beat the average and I'd hope that most folks here would as well.  A bit of education goes a long way!  I'd expect using a pricing engine (pricelabs, wheelhouse, beyond pricing) correctly should put you in the top 25% and beating the numbers airdna has.It's good to do your research, but at some point you just have to make the leap.

John Mark Morelli · Answer

@Ryan Moyer, I thought it might be interesting to the group since AirDNA comes up often on this forum. I don't work for AirDNA, and I am still a bit skeptical of how accurate their algorithm, but here is what they told me when I asked how they differentiated between occupancy and owner-blocks.

"Thank you for contacting AirDNA. I'm pleased you asked how we determine the difference between a reserved day and a blocked day; it's an integral part of our data.We started picking up data on Airbnb in 2014, and in Q4 2015, Airbnb stopped showing the real reservation information. So until then, every day we'd see an A (available for rent and not reserved), R (reserved) and B (blocked), so we knew the exact information of Airbnb reservations.At the end of 2015, they closed that door, and henceforth in the calendar, we can now only see A (available for rent) or U (unavailable), just as the human eye sees. We then built out an algorithm based on the year and a half of historical data to determine which of the Us (unavailable) are Rs (reservation) and which are B (blocked).Our model takes into account 16 different criteria picked up by looking at the historical data set (the length of the booking, the booking lead time, the historical performance of the property) to make that determination. That algorithm is historically correct within a margin of error of 5%. In aggregate, this makes our information extremely accurate in the market, although sometimes on individual properties - if there is some quite unusual behavior, it can get the odd reservation wrong.We have subsequently augmented this information with data partnerships and channel managers who see several hundred thousand reservations. This is the machine learning component of our algorithm that continues to make it more accurate, to adjust for changes in booking behavior over time, for example, and ensures that our algorithm continues to be the most accurate on the market.

"

John Mark Morelli · Answer

@Michael Baum you took my joke before I could get to it ;). More of a paladin myself.

I had a few convos with the customer success team when I was in analysis paralysis prior to purchasing my first and only STR. I am not an expert, but here is what they told me.

Revenue is derived from occupancy and ADR. ADR is not perfect since it's a rolling avg and most guests book vacation properties on higher rate days like weekends and holidays, which will play to your favor.

I believe rates should be determined using the "enemy" strategy, which is referred to commonly elsewhere on this sub.

That leaves occupancy, which is the real magic of their tool. They have an algorithm which supposedly can differentiate between blocked dates (no money, or separate platform) and booked dates.

Not sure how much I trust that really, but I used their occupancy metrics in my financial model to determine whether to buy or skip the locations I evaluated.

And finally, my results. Anecdotally, my bookings through February roughly match or exceed my model, which was based on AirDNA as I mentioned. My area avg booking is 2 1/2 months out, which hits our low season, so I am trying to stack cash until May when it gets hot again.

Happy hunting!

Reese Newell · Answer

@Joe S. If you pull the data from airdna, mashvisor, and airbnb and run a deep statistical analysis to remove a lot of the noise, then they are basically spot on for everything.  If you aren't familiar with how to deep dive in to data and do the analysis to this level then you have 2 options, use it and not know where you are guessing and which sigma bounds your data lies between, or pick up some books on statistics if it is extremely important to you.  The other option is the easiest path of resistance and which most use and just get a semi good idea from them and go from there.

Ryan Moyer · Answer

Interesting.  I never knew that they even tried to parse out owner stays.  I always figured it was part of the expected margin of error.

Most of the markets I am subscribed in when you look at the top properties there are usually one or two that have just insane total revenue despite looking like nothing special as properties.  Like I'm talking 10x average revenue and 5x the revenue of the next closest property.  I always figured those were just properties that the owners used a lot.

Luke Carl · Answer

@Joe S.  It’s about as accurate as it was the last 432289477 times this question was asked ;)

Nick Jacobson · Answer

I take Airdna as one of many data points when comping an investment. We operate in a dense STR area, so there are a ton of data points, but you can not rely on Airdna solely for investment decisions. See what similar properties nearby are asking for on Airbnb/Vrbo and talk to other property managers.

Ryan Moyer · Answer

How do you do that with seasonality involved, though?  For instance I'm looking in Southern Utah right now.  But it's November.  And I know November is completely different than April in Southern Utah, which is completely different than July in Southern Utah.

How accurate is an airDND report?

Most Popular Reply