Good morning, friends. I see a fair amount of posts on here expressing frustration with various issues–faulty bank connections, missing data, etc. I thought it might be useful to have a discussion about how services like Tiller actually work. This is not to let Tiller off the hook for things going wrong, but maybe just to explain how these things break down. I’m not an expert and I’m not in this industry, but I guess I consider finances and technology a bit of a hobby. So, this is based on having played around with a bunch of these services over the years and what I’ve learned:
We start with all of the financial institutions with which we do business. Everything from banks to investment institutions to credit card companies. Each of these institutions maintains data about its customers, and that data may be structured in a variety of different formats. Each of these institutions also has its own security protocols that mandate things like multifactor authentication. The first challenge, then, is to figure out how to get this data from the thousands of financial institutions that we use.
To solve this challenge, services like Tiller contract with what’s called an aggregator. Aggregators, well, aggregate our data from the financial institutions. Tiller’s current aggregator is Yodlee, but there are others such as MX and Plaid (and Tiller has indicated that they’re working on adding a second aggregator). Aggregators gather data from financial institutions through two methods–an Application Programming Interface (API) or what’s called “screen scraping.” An API can be thought of as a key that a financial institution provides to an aggregator to access its data. Screen scraping is where an aggregator essentially deploys a bot with your username and password (and any other authentication that some institutions, but not others, require) to log into your financial institution and retrieve your data. APIs are more secure and easier to use. Scraping is trickier and requires constant adaptation to the changes that institutions make to their web interfaces. Under emerging Open Banking standards, financial institutions would be required to provide APIs that would facilitate access to our financial data for aggregators and other financial institutions.
The key point to understand about aggregators accessing our data is that financial institutions have not always been enthusiastic about aggregators. Nothing is as valuable to financial institutions as the data they have about their customers, and aggregators threaten that. When access to accounts break, it’s often because financial institutions themselves have changed their websites or their security protocols in a way that make it more difficult for aggregators to scrape data from them. Financial institutions do other things that make things more difficult for aggregators like limiting the number of times per day that aggregators (or, more precisely, their bots) can access their data. Services like Tiller can work with their aggregators to improve the service they are contracted to provide, but it’s harder for Tiller (or Yodlee) to compel behavior from financial institutions that don’t want to make it easy to access the data that they control.
Why are some aggregators able to connect to some institutions while others are not? My sense is that it’s often because one aggregator may have figured out a way to scrape data from an institution while another is still figuring it out. Or in some cases, a financial institution may have reached an agreement with an aggregator to access its data through an API that isn’t provided to all of the aggregators. There are thousands of financial institutions out there using all sorts of different standards, so it’s challenging for all of the aggregators to have all of the institutions working all of the time.
After the aggregator has retrieved your data (i.e., transactions and balances), only then does a service like Tiller do its work. Tiller takes the raw data that Yodlee provides and engineers that data so that it will properly fill a Google or Excel spreadsheet. Given the many different formats in which financial institutions store customer data, I suspect (but don’t know for sure) that there can be a “whisper down the lane” challenge whereby Tiller needs to figure out how best to process data that has already gone from the financial institution through Yodlee and then on to Tiller.
To summarize, when you refresh your data, Tiller asks Yodlee to go get your data from your financial institution. Yodlee will then use either an API or scraping to retrieve that data and send it to Tiller. When you fill your sheets, Tiller takes that data from Yodlee, processes it, and makes it available to you.
I think that covers it. Please feel free to correct anything I got wrong as the above is, again, mostly what I’ve pieced together over the years. Not sure if anybody is going to read this brief novel, but I thought it might be helpful, especially to those of you who might be new to these services. The much shorter version:
-Financial institutions have our data.
-Aggregators access that data through either an API or screen scraping.
-Those access methods can break down, especially scraping.
-Tiller (and similar services) process the data that they (hopefully) get from aggregators.
-We go along our merry way managing our finances.
Another thing to keep in mind is that people compare their data access experience to competing products and Tiller understands they need to (and can) improve it, which is why it is their top priority.
And Tiller has an additional challenge to automatically transfer our data into our spreadsheets. Other products effectively stop at the refresh stage, and maybe give users an option to manually download their data - and leave it up to them to convert it to a useable format.
So, Tiller is likely the easiest/fastest way to get our near-real-time data into our own spreadsheets.
Right. Or those competitors download it into their unique app with their unique way of structuring the data with very few options to customize it, which I personally found stifling. As you note, Tiller seems to get it that everybody wants more reliable bank connections. I suspect nobody wants that more than the Tiller team does.
I have tried almost all of the other budget tools. I will take Tiller with issues/bugs over any of them every day of the week. Tiller puts my data in a spreadsheet. Clean. Simple.
No ads, no crappy UI, no other nonsense. That means Tiller can focus on the most important step in this whole journey – aggregating the data for the customer.
I will say that, as @dmetiller mentioned, some financial institutions aren’t friendly to data aggregators. I, personally, stopped using all of those institutions and moved to institutions that do play ball. If they want my business, they need to help make my life easy.
some financial institutions aren’t friendly to data aggregators. I, personally, stopped using all of those institutions and moved to institutions that do play ball. If they want my business, they need to help make my life easy.
This is so important. As an individual, I likely won’t be able to convince a giant institution like Chase Bank to do anything. But if the “pro-aggregator community” bands together, regardless of platform, to tell these banks that we will move our business if they don’t get on board, we can make an impact.
True, though interestingly it’s the big banks that have been first to get on board with Open Banking. Chase, for example, is now following the protocols. My sense is that it’s the smaller credit unions and banks that have been slower to adapt. Not sure we have any more leverage over them, but they probably are more reliant on our business.
I think it’s a bit of a love/hate relationship. Lots of banks/institutions have created “dashboards” in recent years that allow you to see all of your financial data in one place. The banks are relying on an aggregator to create those dashboards. I imagine the business motive is to get you to treat your bank as your one-stop for all things financial. My sense is that most of those dashboards aren’t very good and aren’t widely used, but it does suggest that the banks aren’t antagonistic to the aggregators when it suits them.
My main checking account has been down for nearly a month and isn’t expected to be back online for another three weeks. One of my retirement accounts has been down since December and its not on the alert list and is still on the supported list! I understand the issues and understand that it’s hard work to keep up. But at the same time, I have tried out a different service (the butterfly one) and they have been able to support the accounts that I’ve had issues on, plus have been able to pull transactions on accounts that require 2FA here. I think that business has a few aggregators including plaid. In my opinion Yodlee just doesn’t provide enough support for Tiller to continue its use.
Yeah, I think the key thing is that YMMV with all of this. With literally thousands of financial institutions out there and each of us using some small, unique subset of those institutions, there’s going to be variation in experiences. I don’t know how many customers Tiller has, but I’d be willing to bet that no two of those customers use the exact same institutions.
I’ve never been able to connect to all of my institutions at any of the other services (Monarch, Copilot, Mint, Empower, etc.), but Tiller has been able to do so, which has me a generally happy camper. I fully recognize that’s not everybody’s experience, and I get the grumpiness. As I understand it, Tiller isn’t planning on replacing Yodlee. It’s planning to add another aggregator to supplement it. That strikes me as a sensible plan, and I know at least a couple of those other services have more than one aggregator.
I can imagine the pain of dealing with so many different secure financial institutions. Setting expectations about when a connection will work is Tiller’s responsibility, however. There should be no dates given at all. No timeframe is better than saying 2 weeks and bumping it back every 2 weeks. It’s better to hear that a feature isn’t working now than getting told it’s about to work on a certain date and then it doesn’t.
That’s fair. Fwiw, I’ve never see any of Tiller’s competitors provide the type of transparency that the alerts dashboard does, but there’s a downside to that. I wonder if they’ve ever tracked how accurate their estimates on restoration time are, in general.
Not to fear monger, but since the relationship between aggregators and financial institutions isn’t exactly, “friendly”, is there ever a possibility that relations will break down and we might lose access to our data on a platform like Tiller? Or are there legal obligation that financial institutions must maintain to us, as their consumers?
Good question. I think the good news there is that everything seems to be trending in the right direction. The whole Open Banking movement that I mentioned in my post has picked up considerable momentum, including from people in Congress who have identified the issue as an important one (for whatever that might be worth in the end). Many of the largest banks in the country have moved to Open Banking already. You can generally tell which ones they are on Tiller because they refresh and sync the fastest and smoothest of any of your bank connections.
I’d wager that a decade from now this post will look like ancient history and that aggregators might actually be out of business (i.e., if everybody would provide an API, there may be no need for aggregators to go hunting for our data).
I am interpreting that when you say “we might lost access to our data on a platform like Tiller” you mean that Yodlee (or any other aggregator) might lose access to the data generated by your personal banking/investement accounts. That is a possibility. Your banks/brokerages are obligated to give you access to your data, but they are not obligated to make it easily accessible to third parties.
Depending on the size or other interests of your banks/brokerages, they may decide it is in their financial or organizational interests to restrict access. In that case you are forced to decide if the service they provide is worth the lack of data portability. Nothing gets a monetary organization to change its behavior faster than a loss of customers and I suspect that is what motivates data sharing in the first place - banks/brokerages perceive that their customers (at least a minimal) want their data to be exportable in some format.
If you are committed to using a service like Tiller (which prioritizes keeping the data they pull in your direct hands and centering privacy) you have a part to play in advocating for better data aggregation! Here’s been my experience.
I bank primarily through a credit union and a Vanguard account. Vanguard is a co-op and their technological innovation will be slow rolling - I’m okay with that because I believe in the co-op model and understand they are laser focused on keeping their administrative costs low. Low administrative costs will (at least on a small level) impact customer service robustness, but in the long term helps me build wealth - those are headaches I can bear.
Currently my credit union facilitates data scraping but not APIs. My credit union is not massive, but is probably comparable to a mid-sized regional commercial bank. A quick glance at their organizational tree helped me identify the individual in charge of their tech platform/offerings. I emailed them and articulated the reason I think they should prioritize moving to adopting open banking standards and how it would further incentivize me to use their services. I got a reply from the CTO with some information on their strategic planning around data sharing. In other words - members of their executive team are aware that at least some of their customers want to see these changes implemented.
I’m not expecting then to adopt opening banking overnight - it will probably take half a decade for them to make the necessary IT infrastructure adjustments. In the meantime I’ll put up with some of Tiller’s challenges because I’d rather pay for a service that truly centers the customer over a service that wants access to my data so they can sell that data to larger data brokers or advertisers.
There’s no shame in deciding to use another service! One of the cool things about the way Tiller works is that if you decide to leave you’ve already got all of your data and don’t have to worry about Tiller doing anything with it once you leave because Tiller never took possession of it to begin with.
Well said. Unfortunately, I think it’s the smallest institutions, like many credit unions, that are going to be the slowest to adopt open banking standards. I don’t really know how any of this works or is implemented, but to the extent that adopting these standards is a fixed cost of adopting new technology, that fixed cost is going to hit smaller institutions harder than it will larger institutions.
Yuppp! Overall, I do think the benefits of a highly decentralized banking system are worth the headaches. I’d prefer our model to a Canada or Ireland where there are basically only 3-4 banks.
To get all philosophical for a minute, we’re (U.S.) so habituated into expecting a frictionless service experience that we forget that resilient social institutions (like a banking system, or democratic governance) require some friction and deliberation to operate sustainably.