data on data€¦ · value creation and benefits views on value creation of data and data analytics...

2019 THEMES REPORT

DATA ON DATAHow the agricultural and food business value chain is evolving

2 | Themes Report © 2020 Center for Food and Agricultural Business

Authors

Brady Brewer

Luciano Castro

Nathan DeLay

Scott Downey

Masi Keshavarz

Mati Mohammadi

Contact Information

Center for Food and Agricultural Business Krannert Building, Room 754 403 W. State Street West Lafayette, IN 47907-2056 Phone: (765) 494-4247 E-mail: [email protected] Website: www.agribusiness.purdue.edu

Purdue University Center for Food and Agricultural Business

@PurdueAgBiz

Themes Report | 3 © 2020 Center for Food and Agricultural Business

TABLE OF CONTENTSIntroduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Non-Farmer Survey Results and Insights . . . . . . . 5

Sample Demographics . . . . . . . . . . . . . . . . . . . . . 5

Food Retailers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

Data Collection and Analytics . . . . . . . . . . . . . . 7

Value Creation and Benefits . . . . . . . . . . . . . . . . 8

Data Related Challenges . . . . . . . . . . . . . . . . . . . 9

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Food Manufacturers . . . . . . . . . . . . . . . . . . . . . . . 10



Data Related Challenges . . . . . . . . . . . . . . . . . . . 11

First Handlers/Food Processors . . . . . . . . . . . . 12



Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Ag Input Retailers/Dealers . . . . . . . . . . . . . . . . . 14

Data Usage and Challenges . . . . . . . . . . . . . . . . 14

Comparison with Other Segments . . . . . . . . . . 16

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Ag Input Manufacturers . . . . . . . . . . . . . . . . . . . . 17

Data Collection, Data Use, and Decision Making . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Data Collection and Use Compared to Competitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Collecting Farmer Customer Data . . . . . . . . . . 19

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Farmer Survey Results and Insights . . . . . . . . . . . . 22

Sample Demographics . . . . . . . . . . . . . . . . . . . . . . . . 22

Precision Agriculture Adoption . . . . . . . . . . . . . . . . 22

Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Data Decision Making . . . . . . . . . . . . . . . . . . . . . . . . . 23

Data Management Practices . . . . . . . . . . . . . . . . . . . 24

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Endnotes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27


There is a lot of talk about "big data" in

agriculture these days. The food and agricultural

business value chain is investing in infrastructure

to collect high-quality data and data analytics to

gain key insights into their businesses. The farm

of the future is said to use multispectral imagery,

soil and micro-climate sensors, equipment

telematics data and GPS to drive yield enhancing

decisions. The growth of ag-tech startups suggests

investors are optimistic about this future.

Investment in the ag-tech sector grew 43% in

2018 to nearly $17 billion according to AgFunder

News.1 Though the amount of data being collected

from farms, agribusinesses and food companies

is growing rapidly, little is known about how

different players of the food and agricultural

business value chain leverage this data to make

decisions.

To begin to answer these questions, researchers

from Purdue University have conducted two

surveys: farmer and non-farmer research studies

surveying six levels of the value chain: (1) Food

Retailers, (2) Food Manufacturers, (3) First

Handlers/Food Processors, (4) Farmers, (5) Ag

Retailers and (6) Ag Input Manufacturers.

The non-farmer survey was conducted by Purdue

University's Center for Food and Agricultural

Business from May to July 2019. The purpose of

this study was to explore how organizations in

different segments of the agricultural industry are

acting with regards to collecting, analyzing and

sharing data and to understand how collecting

and sharing data creates value and influences

organizations' decision making. More than 1,800

U.S individuals in different roles responded to

the survey, of which, 1,386 were valid. Survey

participants were asked about their organization's

overall data collection, analytics and value

creation. The questions were then broken down

by business function. Participants were also

asked about data sharing with their customers

or suppliers, the frequency of data sharing and

whether they provide analytics services to their

farmer customers.

The farmer survey conducted by Purdue

University's Center for Commercial Agriculture

surveyed 800 corn and soybean producers

about their collection, management and usage

of farm data. The survey was limited to farms

with 1,000 acres or more to produce a sample of

farms most likely to have active data strategies.

The survey asked respondents 32 questions

regarding farm demographic characteristics,

precision agriculture adoption, types of data

collected on the farm, farm management decisions

influenced by data (if collected), data management

practices and any data sharing with outside

service providers. The goal of the survey was to

understand the farm data lifecycle from collection

to decision making and the various channels

through which data becomes actionable.

INTRODUCTION


Sample Demographics

Figure 1 demonstrates survey participation by

each segment of the food and agricultural

business value chain. Employees working at

agricultural input manufactures had the highest

survey participation of 30%, whereas employees

working at food retailers had the lowest

participation of 12% in this study.

Figure 1. Respondents by Segment in the Food

Value Chain

The survey targeted various job positions and

business functions in order to explore the

perspectives of different job levels and business

functions. As shown in Figure 2, the sample

shows a good representation of executives, upper

managers, middle managers and salespeople.

Figure 3 shows that most of the respondents

were in sales, management, operation and

marketing business functions due to the fact

that the higher number of employees in the

organizations working in these departments or

business functions.

Figure 2. Respondents by Job Position

Figure 3. Respondents by Business Function

The organization size is captured by the

number of employees. Of the 1,358 individuals

represented in this survey, 32% of them work in

an organization with more than 1,000 employees,

and 33% between 100 to 1,000 employees

(Figure 4). Our survey results show the size of

the organization has a positive correlation with

value creation in the firm.

NON-FARMER SURVEY RESULTS AND INSIGHTS


Figure 4. Organization Number of Employees

Figure 5 displays the number of years the

organization has been active in the industry,

representing organization experience. Over 40%

of participants said their organization has been

in the industry for more than 50 years. Only 6%

of the sample represented organizations with less

than 5 years of activity in the industry.

Figure 5. Organization Years in the Industry

Food Retailers

Food retailers are the gatekeepers to the end

consumer along the value chain. Food retailers

are the first to see the products demanded by

end consumers and communicate this message

to suppliers. They also have the largest customer

base and have adapted accordingly to track their

customers' purchasing habits with reward and

loyalty programs. This enables them to better

market and serve their customers. However,

we found that food retailers had some of the

most negative views of their data collection and

analytics efforts.

One hundred and seventy-one respondents

identified as belonging to a company that serves

the food retailer. These respondents represented

a varied background both in terms of role and

job function. Figure 6 shows how respondents

of food retailers were classified by role. Twenty

percent were at the executive level, 33% had direct

reports and identified as a manager or director

within their organization, 15% were salespeople

and 31% were in the "other" category that ranged

from analysts to clerical support.

When examining the demographics by job

function, we see that 39% of respondents

indicated they were most closely related to the

sales function of the business (Figure 7). This

means that some of the executives and managers

were in the sales arm of the firm. After sales, the

largest job functions represented were service at

19%, operations at 10% and human resources at

8%.

This data shows that the majority of respondents

were outward facing in their roles, either

interacting with the customers of the firm or the

suppliers.


Figure 6. Food Retailer Respondents by Job

Position

Figure 7. Food Retailer Respondents by Business

Function

Data Collection and Analytics

One of the first questions we asked was if the

employee's firm collects data. Thirty percent of

food retailers indicated that they collect extensive

data, 48% indicated they collect some data and

22% indicated they collect no data. Food retailers

had the lowest level of data collection compared

to agricultural input manufacturers, agricultural

retailers, processors and first handlers, and food

manufacturers (Figure 29).

When asked about the quality of the data that

is collected, views were fairly similar across all

job functions for food retailers (Figure 8). The

marketing function had the lowest level of high-

quality data (47%) while regulatory compliance

and procurement had the highest (53%). Very

few indicated that the data collected was of poor

quality. This indicates that companies either do

not collect data if it is not of good enough quality,

or they are willing to dedicate enough resources to

ensure it meets a certain standard.

* The definition of data quality was included in the

survey and refers to the completeness, validity,

consistency, timeliness and accuracy that makes data

appropriate for a specific use.

Figure 8. Food Retailers Assess Data Quality by

Business Function

Figure 8 shows that food retailers collect a lot

of data that is able to be utilized since it is of high

quality. However, when we asked respondents to

grade their firm on an A through F scale based

on using data analytics in decision making,

employees of food retailers were not keen to

indicate their firm was doing a great job with the

data. Executives and other high-level employees

were the most positive in their self-assessment

with 42% grading themselves as "A", 33% as "B"

and only 26% as either "C", "D" or "F" (Figure

9). Salespeople were almost as positive as the

executives; however, managers, directors and

other functions such as analysts and clerical were


not as positive. This shows the varying degrees

of impact that data collection and analytics has

across firms.

It is unclear, though, what the driving variation in

the self-assessment results is across the different

job roles. One explanation could be the level of

interaction each role has with the data, such as the

CEO seeing the value creation across the broader

company relative to lower level employees.

Another explanation may be that different

functions have higher expectations of the value

they believe the data can provide. Thus, it is not

necessarily that they think their organization is

doing a bad job with data collection, but that they

believe it could be doing better. Food retailers

graded themselves lower than other areas of the

value chain.

Figure 9. Food Retailers Data Analytics Self-

Assessment by Job Position

Value Creation and Benefits

Views on value creation of data and data analytics

was consistent across the different job functions

(Figure 10). Respondents indicated that data and

data analytics created the least amount of value

now for inbound logistics at only 45.6%. The

customer facing sales function was the highest

with 46.9% indicating it creates value now. This

implies that when a firm starts to collect data and

use analytics, the sophistication and value

opportunities are consistent across the entire

firm. No one particular area is benefiting from the

data as all benefit equally. Another important

point to note is that even for respondents that

indicated no value is being created now, the

majority believe that data and data analytics will

create value in the next five years. Given the level

of investment firms are making in this area,

people expect this investment to pay dividends in

the future.

Figure 10. Food Retailers Assess Value Creation

by Business Function

We know that value is being created now by data

for each job function, but how are these benefits

actually manifesting? Figure 11 shows the

rankings of where respondents believe benefits

will accrue. Note that for this chart, the lower the

number, the better. For example, an item with a

ranking of 3 means more respondents think it is a

bigger benefit than an item with a ranking of 4.5.

All job roles agree that the largest benefit of data

will be improved customer satisfaction. As we

mentioned previously, food retailers by and large

use loyalty or reward systems to track customers.


This allows them to target marketing and

promotional efforts. It is evident that food

retailers believe these data systems help them to

better serve their customers and improve

customer satisfaction. After the benefit to

customer service, respondents believe the biggest

benefit will be to operational efficiency. Given that

most retailers now stock well over 40,000 stock

keeping units (SKU), data may have a large impact

on keeping track and reducing handling costs for

this enormous amount and variety of product.

Figure 11. Food Retailers Assess the Benefits of

Data by Job Position (Average ranking out of 5)

* 1 is "the most important" benefit and 5 is "the least important"

benefit

Data Related Challenges

Lastly, we asked food retailer respondents

about the challenges to data (Figure 12). All job

roles agree that security and privacy concerns

is the biggest data challenge for food retailers.

It should be noted that food retailers were

unique in selecting security and privacy as the

biggest challenge. Other areas of the food value

chain indicated that timeliness was the biggest

challenge. At the risk of sounding like a broken

record, food retailers have a mountain of data

coming in by the minute as people shop with

loyalty and reward systems. However, some

high-profile security breaches at large national

retailers have put security and privacy concerns

on the minds of retail customers. It is apparent

that food retailers have heard the concerns of their

customers.

Figure 12. Food Retailers Assess Data Related

Challenges by Job Position (Average scale out of

10)

*Where 1 is "not important" and 10 is "very important"

Conclusion

Food retailers are uniquely positioned as they

reach millions of United States consumers on

a daily basis. While it may seem that they have

a large amount of data to use, food retailer

respondents were the most negative when it came

to data collection and the self-assessment of their

firm. This may be attributable to food retailers

being at a different stage in the data lifecycle as

the rest of the food value chain as many food

retailers interact with many different industries.

Regardless of their definition of data and how they

view themselves, it is apparent that they believe

they can do better in regard to data collection,

analytics and the value they drive.


Food Manufacturers

Food manufactures take raw or processed

inputs and transform them into a commercially

viable products to be sold in grocery stores and

restaurants. Food manufactures are helping

shape the food value chain by manufacturing

the products that consumers demand. Whether

it is a product that is organic or non-GMO, food

manufactures are developing production methods

to satisfy these demands. At the same time, food

manufacturers are also satisfying traceability

and food safety requirements that food retailers

and consumers alike are making an industry

standard.

The majority of respondents from the food

manufacturer group were either executive level

employees or individuals that have a direct

report (manager/director). The distribution of

respondents is shown in Figure 13.

Figure 13. Food Manufacturer Respondents by

Job Position


Food manufactures collect the second most

amount of data in the agricultural value

chain, only trailing behind agricultural input

manufacturers. Forty-seven percent of food

manufacturers indicated that their company

collects extensive data. Another 43% indicated

that their company collects some data and only

10% indicated that their company collects no data

(Figure 29).

Food manufacture respondents were fairly

consistent with their answers about the quality of

data being collected (Figure 14). Most business

functions believe more that 50% of data being

collected is high quality. Human resource

management, procurement and the service

functions reported the highest quality of data

while sales, inbound logistics and marketing

reported the lowest quality of data.

Figure 14. Food Manufacturers Assess Data

Quality by Business Function

Next, we asked respondents to grade their

company on data and analytics. Executives and

employees who have a direct report were the most

optimistic about their company's data activities

(Figure 15). Thirty-nine percent of executives in

food manufacturing firms that responded to our

survey graded themselves as an "A", while 35%

graded their company as a "B". Sales employees

were the most negative toward their company's

data activities. Sixteen percent of sales employees

graded their company as an "A", while 28%


graded their company as a "B". Additionally, 24%

of sales employees graded their company as an

"F". Only 2% of executives graded their company

as an "F".

Figure 15. Food Manufactures Data Analytics

Self-Assessment by Job Position


Views on value creation of the data and data

analytics was consistent across all job functions

(Figure 16). Respondents indicated that data

creates the least amount of value now for

procurement and operations and creates the most

value for the service and sales function of the

business.

Figure 17 shows food manufacturers' views on

the biggest benefits of data. It is important to note

that the lower the score, the bigger the potential

benefit. Food manufacturers, like most other

parts of the agricultural value chain, indicated

that improved customer satisfaction is the biggest

potential benefit to data and analytics. It should be

noted that sales employees also ranked improved

operational efficiency as high as improved

customer satisfaction. Improved compliance and

market awareness were indicated as benefiting

the least from data.

Figure 16. Food Manufacturers Assess Value

Creation by Business Function

Figure 17. Food Manufacturers Assess the

Benefits of Data by Job Position (Average ranking

out of 5)


benefit

Data Related Challenges

When asked about the challenges of collecting and

analyzing data, employees of food manufacturers

did not agree (Figure 18 and 19). Executives

indicated that timeliness and availability are the

biggest challenges. Sales employees indicated that

analysis technology and talent are the biggest

challenges. Employees with direct reports

(managers/directors) who are not executives

ranked every obstacle as a bigger challenge than


their fellow food manufacturing employees.

Overall, it seems that company leadership views

the process of acquiring data and turning it into

insights in a timely manner as large challenges

while lower level employees — those collecting the

data and analyzing it — view the resources they

have as the biggest challenge.

Figure 18. Food Retailers Assess Data Challenges

by Job Position (Average scale out of 10)


Figure 19. Food Retailers Assess Data Challenges

by Job Position, Continued

First Handlers/Food Processors

First handlers and food processors are the first to

touch the agricultural good after it leaves the farm

gate. As such, this part of the food value chain

buys a large amount of product from farmers and

sees upstream demand from food manufacturers

and food retailers. They are responsible for safe

handling and taking homogenous commodities

and turning them into different value streams to

be made into end products across the world.

Two hundred and twenty-nine respondents

belonged to the first handler and food processor

group (Figure 20). The demographics of the

respondents are as follows: 86 at the executive

level, 82 with direct reports, 33 in sales and 28 as

"other", which represents analysts, clerical and

other business roles.

Figure 20. First Handlers/Food Processors

Respondents by Job Position


First handlers and food processors were in the

middle in terms of data collection compared to

other parts of the food value chain. Forty-one

percent of first handlers and food processors said

their firm collects extensive data, 51% indicated

they collect some data and 8% indicated they

collect no data. However, one area where this part

of the food value chain excelled is in the collection

of high-quality data. Fifty-eight percent of

respondents said that the data their firm collects

is high quality, and 36% is of medium quality. This

is the highest quality relative to what respondents

of other areas of the value chain indicated.


However, despite having high-quality data, when

asked to do a self-assessment of their firm's data

analytics efforts, first handlers and food

processors had a negative view (Figure 21).

Executive level employees were the most negative

of their firm's data efforts with 17% of

respondents grading themselves as "A", 42% as

"B", 32% as "C" and 9% as "D". Sales staff were

the most positive in the self-assessment with 35%

grading themselves as "A", 35% as "B", 19% as "C"

and 8% as "D".

Figure 21. First Handlers/Food Processors Data

Analytics Self-Assessment by Job Position


When asked about data creating value for their

firm, 44% of first handlers and food processors

indicated it creates value now — the lowest level

of value creation in the survey. Thirty-one percent

of respondents from the first handler and food

processor group indicated that they expect it

to create value in five years. This level of value

creation is contradictory with the quality of data

that this level of the food value chain indicated

they collect. The high-quality data is not turning

into realized value for these firms.

Next, we asked about the benefits of the data

(Figure 22). First handlers and food processors

indicated that improved customer satisfaction is

the largest benefit they see in the data. This

answer is not unique to first handlers and food

processors and was a common response across

the value chain. However, the second biggest

benefit is more unique as first handlers and food

processors indicated market awareness as the

next biggest benefit. Put in the context of the

competitive commodity markets these firms

operate in, this choice of benefit is not surprising.


Assess the Benefits of Data by Job Position

(Average ranking out of 5)


benefit

When asked about the biggest challenge to data,

first handlers and food processors indicated

timeliness as the number one challenge (Figures

23 and 24). Timeliness was a common theme

across the value chain as data can only help if it is

relevant and up-to-date.



Assess Data Challenges by Job Position (Average

scale out of 10)



Assess Data Challenges by Job Position,

Continued

Conclusion

Overall, first handlers and food processors

see a large amount of data on a daily basis as

commodity markets change by the second. First

handlers take commodities and goods from

farmers, split them into different value streams

and distribute these goods across the globe.

Ag Retailers/Dealers

Themes that emerged from the ag retailers

survey results fell into three categories: 1) there

are differences in perception about data between

functions; 2) ag retailers' and dealers' hold may be

insulated, such that their comparative views with

other parts of the value chain are unique; and 3)

there may be some interesting dynamics as ag

retailers/dealers work through these issues in the

future.

Data Usage and Challenges

To understand different perspectives within

ag retailer/dealer organizations, respondents

were grouped into four categories: executives,

managers, salespeople and other. Managers

represented the largest group of respondents

(n=145), executives were second largest (n=114),

salespeople were third (n=69) and the "other"

category (made up of analysts, administrators and

others, n=32) was the smallest.

Overall, ag retailers grade themselves with

regards to their organization's data analytics

efforts fairly consistently across all functions with

4% (salespeople) to 8% (executives and other) of

respondents in each category grading their

organization's use of data as an "A". Close to half

of all respondents graded their organization at a

"B", and close to a third gave their organization a

"C" (Figure 25); however, these overall grades did

not represent grades given to all business

functions. There were marked differences

between executives and salespeople.


Figure 25. Ag Retailers Data Analytics Self-

Assessment by Job Position

While views of the HR function were similar to

the breakdown of overall grades for the

organization, the use of data in procurement and

logistics tended to be more positively viewed,

particularly by executives. About 15% of

executives gave these areas an "A". Salespeople

were less positive about inbound logistics with

none giving their organization an "A" for this

function. Salespeople were more positive about

their organization's data use for compliance with

24% of them grading this area as "A", compared

to only 14% of executives. Salespeople were also

more positive about grades for sales and

marketing than executives, although data usage

for service showed executives as markedly more

positive (Figures 26 and 27).


Evaluation by Position – Sales Business Function


Evaluation by Position – Service Business

Function

There may be different perspectives on how

good use of data is defined at various levels of

the organization. It is possible that salespeople

view data collection as "usage", while managers

and executives may define usage as analysis and

insights to make decisions. These differences

are reflected in responses about whether data

has more potential to be used in the future as

compared to its use today. Forty six percent of

executives responded that there would be more

value from marketing data five years from now

than there is today. The majority of people in all

functions in retail organizations felt that sales

data was valuable today, but just over a third of

salespeople felt that service data would be more

valuable in five years.

While everyone stated that timeliness of data was

one of the largest challenges, other perspectives

were split functionally. Salespeople are asked to

collect a large amount of data. They reported that

the technology used to collect and analyze this

data is their second biggest challenge. Managers

and others who analyze data report that the talent

and skill to collect and analyze it are lacking.

While executives state infrastructure is a lesser


challenge, managers and salespeople are less

convinced that this is true (Table 1).

In summary, within retailer organizations,

perspectives on data are nuanced depending on

the position within the organization.

Table 1. Data Related Challenges by Job Position

(1 is "not important" and 10 is "very important")

Comparison with Other Segments of the Value Chain

Compared to other parts of the value chain, ag

retailers/dealers are not as positive about their

data usage as compared to competitors. Still,

about half of ag retailers/dealers believe they are

ahead of the competition in both data collection

and analysis (Figure 28).

Figure 28: Data Collection Compared to

Competitors

While only 30% of food retailers say that their

organizations collect extensive data, 39% of ag

retailers/dealers say they collect an extensive

amount (Figure 29). However, objectively, Point-

Of-Sale data, shopper surveys, shelf-space

analytics and numerous other sources of data

collected at most food retailing organizations

today are far more extensive than most done by ag

retailers/dealers. It is important to remember that

"extensive" is in the eye of the beholder.

Figure 29: Data Collection by Segment

Although there are not big differences in

responses from any part of the value chain, more

than 53% of ag retailer/dealers report that they

use data to make decisions less than half the time

(Figure 30). This is less usage of data to make

decisions than any other part of the value chain.

Ag retailers/dealers seem to be aware of this,

however, as only 6% graded themselves with an

"A" on their use of data to make decisions. Part of

the reason for this, though, may be tied to the

quality of the data. Ag retailers/dealers had the

fewest respondents (28%) stating they had high-

quality data as compared to other parts of the

value chain (Figure 31). However, 58% of ag

retailer/dealer respondents said the data that

exists is used to create value today, the second

highest percentage of the five parts of the value

chain that were examined.


Figure 30: Percentage of Decisions Made Based

on Data Analytics by Segment

Figure 31. Data Quality by Segment

Conclusion

There are a couple of interesting takeaways

from ag retailer/dealers' responses. First, the

respondents who see themselves as far ahead

on data usage rate the benefits of "Improved

Market Awareness" higher than all other benefits

they receive. In contrast, respondents who see

themselves as far behind rate the benefits of

"Improved Compliance with Data Protection and

Privacy Regulations" more highly than others.

Seeing the market value of data seems to correlate

with effort to use it, or at least perceptions of

using it.

There are also differences between these two

groups in terms of the challenges they see.

While both groups rank timeliness as one of the

biggest challenges, ag retailers/dealers who see

themselves as far ahead tend to see challenges

of data availability, security, privacy, volume and

ownership as larger challenges than those who

see themselves as far behind. In contrast, those

who report being far behind are more concerned

about talent, costs and infrastructure.

As these two diverse groups of ag retailers/dealers

look to the future, they will no doubt need to

be clear about the purpose of data within their

organizations and the cost/benefit analysis of

overcoming these challenges.

Ag Input Manufacturers

Ag input manufactures have played a major role

in bringing digital platforms for data collection

and analysis to the table. Three of the most used

farm management software systems come from

manufacturers (Climate, JD center and Case IH),

according to the survey.

Today, most — if not all — major ag input

manufacturers have put together a digital solution

for their organization. These solutions

complement each other, but also compete. There is

a clear a race in the market to present tools that

farmers should adopt (Figure 32).

Figure 32: Selected Digital Agricultural

Initiatives by Ag Input Manufacturers


Data Collection, Data Use, and Decision Making

As shown in Figure 33 and 34, the majority of

survey respondents occupy upper management

and director positions and have a hand in sales

and marketing functions. This group has a high

level of understanding of their organization and

its exposure to external forces such as competitors

and customers.

Figure 33: Ag Input Manufacturer Respondents

by Role/Position

Figure 34: Ag Input Manufacturer Respondents

by Business Function

Figure 29 shows that out of the five agri-

food value chain levels surveyed, ag input

manufacturers have the highest perception

of collecting extensive data. It is worth noting

that although ag retailers are closer to farmers,

manufacturers seem to be more dedicated to data

collection.

When we broke down the question of how

extensively ag input manufacturers were

collecting data by business function (not shown

here), high scores were shown in the areas of

operations, sales, marketing, regulatory and

compliance, while services and HR showed much

lower scores.

Similar results are observed for the question of

the degree to which data is used for analysis and

decision making. Ag input manufacturers have

the highest perception of extensively analyzing

data to make decisions, as shown in Figure 30.

While Figure 29 shows that ag input

manufacturers consider their organizations to be

data-driven, Figure 35 shows that they hold high

expectations for themselves in regard to how well

their organizations use data for decision making.

Ag input manufacturers graded themselves with

the second to lowest "A" score of 7%, indicating

that they feel there is room for improvement in

how they are using data to make decisions. This

indicates that input manufacturers and dealers

have a similar perception on how effectively they

use data for decision making. This may indicate

that interacting with farmers and using data

extensively is a challenge for them. One may also

assume that these two upstream value chain levels

suffer from hurdles related to the challenging

nature of farming (exposure to uncontrolled

variables), while the downstream levels — food

manufacturers, food retailers and food processors

— perceive themselves as more effectively using

data analysis, maybe due to their more traditional

industrial B2B roles.


Figure 35: Data Analytics Self-Evaluation by

Segment

Data Collection and Use Compared to Competitors

The survey also asked each value chain level how

their organization's data collection and use

compared to that of their competitor's. In

comparison to the other levels of the value chain,

ag input manufacturers ranked highest (59%) in

believing they are at par or somewhere behind the

competition in data collection (Figure 36).

Similar results are seen in Figure 37 for data

analytics. These results lead us to wonder if

everyone is struggling, or if there are clear leaders

in every ag input manufacturing segment that are

giving everyone else the perception that they are

somehow behind.

Figure 36: Data Collection Compared to

Competitors

Figure 37: Data Analytics Compared to

Competitors

Collecting Farmer Customer Data

The survey asked the ag input retailers and ag

input manufacturers if they are collecting and

using farmer customers' input and output data.

This type of data can explore what product or

service a farmer applied or used and how it

performed in comparison to the competition.

As can be seen in Figure 38, ag input

manufacturers are not collecting farmers'

input data as much as ag retailers (23.5% for ag

input manufacturers compared to 33.1% for ag

retailers), but neither group is exploring much

farmer output data (both around 20%). This is

a remarkable finding as farmers' output data

would help ag input manufacturers showcase and

prove the value proposition of their products and

services to farmers. This should be a critical use

for collected data in this area.


Figure 38: Collection of Farmer Customers'

Input and Output Data

Some ag input manufacturers are coordinating

with farmers to capture data and use it across the

value chain to generate value. An example from

Syngenta is a trading platform called Nucoffee

(Figure 39, upper part). Syngenta Brazil works

with coffee farmers to help them improve quality

by providing inputs and services. In return,

Syngenta uses grower information regarding

production to leverage demand for coffee

produced.

Another example from Syngenta U.S. is the

partnership between The Nature Conservancy,

ag retailers (such as Nutrien and others)

and Kellogg's cereal company to implement

sustainable agricultural practices and track

the results of such implementation in terms of

sustainability, yields and profitability using its

digital ag tool, AgriEdge.

Third, Bayer invested in a trading platform called

Made in Farm (Figure 39, lower part). It is a

coffee marketplace from coffee growers to roasters

and end consumers that sells coffee from Bayer

customers. It also provides coffee growers inputs

and services to help them grow and sell high-

quality coffee.

Figure 39: Food Value Chain Initiatives by Ag

Input Manufacturers

When it comes to the use of farmer information

(input and output data), we still see limited use to

support input recommendation. Only a little over

40% of the sample perceive their companies as

using this data all or most of the time (Figure

40). As "trusted advisors" should base their

recommendations on factual information, we saw

this as an opportunity for improvement.

Figure 40: Frequency of Using Farmer

Customers' Data to Provide Support

In Figure 41, we see that data storage is not

perceived as a challenge. Over 60% of ag input

manufacturers have their own platform for data

storage, and 30% outsource to a third party.


Figure 41: Platforms Used to Store Farmer

Customers Data - Ag input Manufacturers

Conclusion

There are several interesting takeaways from ag

input manufacturing results that are listed below:

• Digital agriculture solutions are appearing

from all major ag input manufacturers

• Ag input manufacturers are collecting and

analyzing data extensively

• Ag input manufacturers have the highest

expectations to better use data to support their

decisions and the decisions of their customers

• Within different segments of manufacturers,

there is a perception that competitors are

doing better (everyone is struggling and/or

there is a clear leader for every segment)

• Timeliness, availability and people are the

biggest challenges

• The use of input and, mainly, output farmer

information has clear room for improvement

• The use of farmers' data to support product

value propositions could be higher

• Industry boundaries are being challenged

by the need for a more integrated view on

farming

• Data analytics are clearly seen as a new

capability for ag input manufacturers


Sample Demographics

Sample farm characteristics are displayed in

Table 2. Fifty percent of farms surveyed operate

between 1,000 and 1,999 acres, while 36%

are between 2,000 and 4,999 acres and 15%

haver 5,000 acres or more. The sample is more

representative of large commercial operations

(by design). About 80% of surveyed farms have

owner/operators over the age of 50 and 35%

are over the age of 65. Given the average age of

producers in the U.S. is 59, our sample is generally

representative of the national age distribution

of farmers.2 Just under half of sampled farms

have a bachelor's degree or higher as the

highest educational attainment among full-time

employees and 9% have a post-graduate degree

(Master's degree and up), indicating a high degree

of human capital.

Precision Agriculture Adoption

High speed internet access is slightly more

available among survey respondents than rural

America as a whole at 80%. Generally, adoption

rates of precision agriculture technologies reflect

the large commercial size and crop mix of farms in

the sample. GPS guidance or auto-steer for farm

equipment is used by over 90% of the surveyed

farms. Fifty-nine percent of farms use variable

rate technology (VRT) for seeding and 71% use

VRT for fertilizer application. Drones or

unmanned aerial vehicles (UAVs) are used by 26%

of sampled farms. The survey sample has

significantly higher rates of precision agriculture

adoption than the most recent estimate from the

USDA Agricultural Resource Management

Survey (ARMS) (see Schimmelpfennig, 2016) but

are highly similar to work by Thompson et al.

(2018) who use a similar sampling method.3

Table 2. Farm Demographics & Precision

Agriculture Adoptiona

Notes: a Survey sample includes 800 corn and soybean farms

with 1,000 acres or more in operation. b Highest level of

educational attainment among all full-time employees of the

farm, including owner/operators.

Data Collection

Collection among the sample is common —

particularly for yield monitor and soil data at

82% and 77%, respectively. Satellite or drone

imagery data is least likely to be collected (47%

of the sample), but given the novelty of the

technology, this could be considered high. The

FARMER SURVEY RESULTS AND INSIGHTS


vast majority (73%) create GPS maps from their

data, suggesting a high degree of engagement with

data once collected.

Data collection is strongly related to farm

characteristics. Figure 42 displays the percentage

of farms collecting each data type by farm size.

Data collection is most prevalent among large

farms — a result consistent with previous

findings. Collecting farm data — particularly

drone or satellite imagery data — likely involves

scale economies that favor larger operations.

Figure 42. Types of Data Collected by Farm Size

(Acres)

Farms with young operators and more

educational attainment are generally more prone

to collecting farm data. Again, imagery data from

a drone or satellite bears the clearest relationship

to age and education. Of the 800 respondents, 58

(7%) do not collect any of the data types included

in the survey. When asked to identify the primary

reason for not collecting farm data, 36% said

data collection is "too costly", while 19% find the

benefits of doing so unclear. Taken together, over

half of non-collectors perceive farm data to be

un-profitable. Over one-third report uncertainty

in how to use farm data once collected, suggesting

a disconnect between collection and action.

Surprisingly, only 10% of farms cited privacy

concerns as the reason for not collecting farm

data.

Data Decision Making

Farmers that currently collect data were asked to

rate the extent to which their data influences their

decision making in three crop management areas:

seeding rates, nutrient management/fertilizer

application and drainage investments. Figure 43

summarizes the responses. Farm data appears to

have the largest influence on nutrient

management with 93% reporting their fertilizer

decisions to be "somewhat" or "highly" influenced

by data. The share of farms reporting seeding rate

and drainage decisions as at least somewhat

influenced by data is 81% and 71%, respectively.

Fertilizer application decisions are nearly twice as

likely to be "highly" influenced by farm data as

seeding rate and drainage investment decisions,

reflecting the popularity of variable rate fertilizer

application within the sample.

Figure 43. Management Decisions Influenced by

Data

Farms making decisions based on their data

appear satisfied with the results. Seventy-two

percent of those making data-driven seeding


rate decisions report a positive yield impact

vs. 81% for fertilizer decisions and 85% for

drainage decisions. Levels of satisfaction rise as

farmers collect more data types. For example, the

proportion indicating a positive yield result from

data-informed seeding rate decisions is 64% if

the farm only collects one type of data (e.g. just

yield monitor data), but rises to 77% if the farm

collects all three data types — a 21% increase. This

suggests that individual data streams are made

more actionable when combined with other data

sources.

Data Management Practices

The survey broadly focuses on two data

management practices in the farm data pipeline:

adoption of farm data software platforms and

sharing of data with outside service providers.

Overall, 47% of farms that collect data use at least

one data software product, but adoption rates are

significantly higher among larger operations —

63% of farms with 5,000 acres or more vs. 36% of

farms in the 1,000-1,999-acre category.

Farms with higher educational attainment have

higher rates of farm data software adoption, but

the relationship varies by operator age. Figure

44 shows that, among operators over 65, those

with some college are nearly twice as likely to

use farm data software than those with a high

school diploma. Getting a bachelor's degree has

a similar effect on adoption rates among those

age 51-65. Software platforms are popular with

young operators across all levels of education, but

adoption rises to nearly 70% for those with a post-

graduate degree (e.g. master's or Ph.D.).

Farmers that use at least one data service platform

were asked to identify all of the products they

currently subscribe to from a list of eight popular

brands (Figure 45). The most widely used

software product is Climate FieldView (Bayer),

used by over half of surveyed software

subscribers. Forty-four percent use John Deere

Operations Center, while 22% use Case IH's AFS

Software platform — generally reflecting their

respective market shares for farm equipment.

Trimble is the next most frequently used at 21%,

followed by Farmers Business Network (FBN)

(19%), Corteva's Encirca (14%), FarmersEdge

(10%) and Granular (also Corteva) at 9%. Nearly

one-fourth of users subscribe to a service not

listed in the survey, suggesting a long tail in the

farm data software market.

Figure 44. Use of Farm Data Software by

Education and Age

Figure 45. Types of Software Used by Software

Users (n = 353)


Our survey indicates that 63% of subscribers

receive seed or fertilizer application

recommendations (prescriptions) from their

software. However, only 44% follow their

software recommendations "closely", while 52%

follow "somewhat closely" and 4% do not follow

recommendations at all.

Seventy percent of software users subscribe to

more than one product, indicating the absence

of a "one-stop-shop" for farm data solutions.

Although farms use an average of two software

platforms, almost 90% subscribe to three or

fewer. This implies an upper-bound on the

number of platforms farms are willing to adopt.

Given the growth of investment in farm-facing

technology companies, it may be difficult to incent

existing adopters to add another product to their

software suite. Companies could instead target

non-adopters. Of farmers not using any farm

data software, close to half indicate uncertainty

in how to use the technology as the primary

reason for not subscribing. Forty-one percent of

non-adopters perceive farm data software as too

costly or that the associated benefits are unclear,

indicating a breakdown in value proposition.

Privacy concerns are surprisingly unimportant as

a deterrent to software use. Only 12% identified

privacy as the main reason for not subscribing to a

farm data service.

Farmers were asked if they share their data

with agronomists, agricultural input suppliers

and equipment dealers/manufacturers for

the purpose of generating crop management

recommendations. Over 70% of respondents

share their farm data with at least one service

provider, and of these, 63% share with two or

more. Farmers are most willing to share data with

service providers that operate close to on-farm

crop management decisions. Fifty-eight percent of

farms share data with their agronomist followed

by ag input suppliers at 44%. Only 12% report

sharing their data with equipment dealers and

7% share with a service provider not listed in the

survey.

Surprisingly, the share of farms that follow

recommendations provided by outside service

providers "very closely" is 31%, 13 percentage

points lower than the share that follow their

software recommendations closely. The

willingness to follow software generated

recommendations over those provided by an

outside advisor may be due to a perception

that service providers — particularly ag input

suppliers — pair recommendations with product

sales.

We find that farm data is more influential in

the crop management decisions of farms that

subscribe to a data software platform or share

their data with an outside service provider. Farms

that perform both data management practices are

over four times more likely to make seeding rate

decisions that are "highly influenced" by data than

farms that collect data but do not share or use

software. The proportion of data-driven fertilizer

decisions among the software plus sharing cohort

is over twice that of the no-software, no-sharing

group. Software use is positively associated with

the degree to which data influences drainage

investments. Data sharing, however, has little to

no impact on the importance of data in drainage

decisions. Similarly, farms that subscribe to a

data software product are more likely to report


increased yields as are farms that share data with

an agronomist, input supplier, dealer or other

service provider.

Conclusion

Often, ag data ends up siloed on the farm, stored

on desktops or USB flash drives collecting dust

in a shop drawer. These survey results suggest

that proactively managing and analyzing farm

data can improve decision making and generate

positive yield results. Understanding how data

management practices shape on-farm decision

making is of crucial importance in bringing

the farm of the future into reality. Downstream

players in the agricultural value chain must

recognize the data needs of growers as data

transparency and data integration demands rise.


1. AgFunder. AgFunder AgriFood Tech Investing Report: 2018 Year in Review. (2019). Available at https:// agfunder.com/research/agrifood-tech-investing-report-2018/

2. 2017 USDA Census of Agriculture

3. Schimmelpfennig, D. (2016). Farm Profits and Adoption of Precision Agriculture. U.S. Department of Agri-culture, Economic Research Service Report No. 217, Washington, D.C.; Thompson, N. M., C. Bir, D. A. Widmar, J. R. Mintert. (2018). Farmer Perceptions of Precision Agriculture Technology Benefits. Journal of Agricultural and Applied Economics 51(1): 142-163.

ENDNOTES

WWW.AGRIBUSINESS.PURDUE.EDU

data on data€¦ · value creation and benefits views on value creation of data and data analytics...

Documents